Initial commit

models/tf.py

+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+"""
+TensorFlow, Keras and TFLite versions of YOLOv5
+Authored by https://github.com/zldrobit in PR https://github.com/ultralytics/yolov5/pull/1127
+
+Usage:
+    $ python models/tf.py --weights yolov5s.pt
+
+Export:
+    $ python export.py --weights yolov5s.pt --include saved_model pb tflite tfjs
+"""
+
+import argparse
+import sys
+from copy import deepcopy
+from pathlib import Path
+
+FILE = Path(__file__).resolve()
+ROOT = FILE.parents[1]  # YOLOv5 root directory
+if str(ROOT) not in sys.path:
+    sys.path.append(str(ROOT))  # add ROOT to PATH
+# ROOT = ROOT.relative_to(Path.cwd())  # relative
+
+import numpy as np
+import tensorflow as tf
+import torch
+import torch.nn as nn
+from tensorflow import keras
+
+from models.common import (C3, SPP, SPPF, Bottleneck, BottleneckCSP, C3x, Concat, Conv, CrossConv, DWConv,
+                           DWConvTranspose2d, Focus, autopad)
+from models.experimental import MixConv2d, attempt_load
+from models.yolo import Detect, Segment
+from utils.activations import SiLU
+from utils.general import LOGGER, make_divisible, print_args
+
+
+class TFBN(keras.layers.Layer):
+    # TensorFlow BatchNormalization wrapper
+    def __init__(self, w=None):
+        super().__init__()
+        self.bn = keras.layers.BatchNormalization(
+            beta_initializer=keras.initializers.Constant(w.bias.numpy()),
+            gamma_initializer=keras.initializers.Constant(w.weight.numpy()),
+            moving_mean_initializer=keras.initializers.Constant(w.running_mean.numpy()),
+            moving_variance_initializer=keras.initializers.Constant(w.running_var.numpy()),
+            epsilon=w.eps)
+
+    def call(self, inputs):
+        return self.bn(inputs)
+
+
+class TFPad(keras.layers.Layer):
+    # Pad inputs in spatial dimensions 1 and 2
+    def __init__(self, pad):
+        super().__init__()
+        if isinstance(pad, int):
+            self.pad = tf.constant([[0, 0], [pad, pad], [pad, pad], [0, 0]])
+        else:  # tuple/list
+            self.pad = tf.constant([[0, 0], [pad[0], pad[0]], [pad[1], pad[1]], [0, 0]])
+
+    def call(self, inputs):
+        return tf.pad(inputs, self.pad, mode='constant', constant_values=0)
+
+
+class TFConv(keras.layers.Layer):
+    # Standard convolution
+    def __init__(self, c1, c2, k=1, s=1, p=None, g=1, act=True, w=None):
+        # ch_in, ch_out, weights, kernel, stride, padding, groups
+        super().__init__()
+        assert g == 1, "TF v2.2 Conv2D does not support 'groups' argument"
+        # TensorFlow convolution padding is inconsistent with PyTorch (e.g. k=3 s=2 'SAME' padding)
+        # see https://stackoverflow.com/questions/52975843/comparing-conv2d-with-padding-between-tensorflow-and-pytorch
+        conv = keras.layers.Conv2D(
+            filters=c2,
+            kernel_size=k,
+            strides=s,
+            padding='SAME' if s == 1 else 'VALID',
+            use_bias=not hasattr(w, 'bn'),
+            kernel_initializer=keras.initializers.Constant(w.conv.weight.permute(2, 3, 1, 0).numpy()),
+            bias_initializer='zeros' if hasattr(w, 'bn') else keras.initializers.Constant(w.conv.bias.numpy()))
+        self.conv = conv if s == 1 else keras.Sequential([TFPad(autopad(k, p)), conv])
+        self.bn = TFBN(w.bn) if hasattr(w, 'bn') else tf.identity
+        self.act = activations(w.act) if act else tf.identity
+
+    def call(self, inputs):
+        return self.act(self.bn(self.conv(inputs)))
+
+
+class TFDWConv(keras.layers.Layer):
+    # Depthwise convolution
+    def __init__(self, c1, c2, k=1, s=1, p=None, act=True, w=None):
+        # ch_in, ch_out, weights, kernel, stride, padding, groups
+        super().__init__()
+        assert c2 % c1 == 0, f'TFDWConv() output={c2} must be a multiple of input={c1} channels'
+        conv = keras.layers.DepthwiseConv2D(
+            kernel_size=k,
+            depth_multiplier=c2 // c1,
+            strides=s,
+            padding='SAME' if s == 1 else 'VALID',
+            use_bias=not hasattr(w, 'bn'),
+            depthwise_initializer=keras.initializers.Constant(w.conv.weight.permute(2, 3, 1, 0).numpy()),
+            bias_initializer='zeros' if hasattr(w, 'bn') else keras.initializers.Constant(w.conv.bias.numpy()))
+        self.conv = conv if s == 1 else keras.Sequential([TFPad(autopad(k, p)), conv])
+        self.bn = TFBN(w.bn) if hasattr(w, 'bn') else tf.identity
+        self.act = activations(w.act) if act else tf.identity
+
+    def call(self, inputs):
+        return self.act(self.bn(self.conv(inputs)))
+
+
+class TFDWConvTranspose2d(keras.layers.Layer):
+    # Depthwise ConvTranspose2d
+    def __init__(self, c1, c2, k=1, s=1, p1=0, p2=0, w=None):
+        # ch_in, ch_out, weights, kernel, stride, padding, groups
+        super().__init__()
+        assert c1 == c2, f'TFDWConv() output={c2} must be equal to input={c1} channels'
+        assert k == 4 and p1 == 1, 'TFDWConv() only valid for k=4 and p1=1'
+        weight, bias = w.weight.permute(2, 3, 1, 0).numpy(), w.bias.numpy()
+        self.c1 = c1
+        self.conv = [
+            keras.layers.Conv2DTranspose(filters=1,
+                                         kernel_size=k,
+                                         strides=s,
+                                         padding='VALID',
+                                         output_padding=p2,
+                                         use_bias=True,
+                                         kernel_initializer=keras.initializers.Constant(weight[..., i:i + 1]),
+                                         bias_initializer=keras.initializers.Constant(bias[i])) for i in range(c1)]
+
+    def call(self, inputs):
+        return tf.concat([m(x) for m, x in zip(self.conv, tf.split(inputs, self.c1, 3))], 3)[:, 1:-1, 1:-1]
+
+
+class TFFocus(keras.layers.Layer):
+    # Focus wh information into c-space
+    def __init__(self, c1, c2, k=1, s=1, p=None, g=1, act=True, w=None):
+        # ch_in, ch_out, kernel, stride, padding, groups
+        super().__init__()
+        self.conv = TFConv(c1 * 4, c2, k, s, p, g, act, w.conv)
+
+    def call(self, inputs):  # x(b,w,h,c) -> y(b,w/2,h/2,4c)
+        # inputs = inputs / 255  # normalize 0-255 to 0-1
+        inputs = [inputs[:, ::2, ::2, :], inputs[:, 1::2, ::2, :], inputs[:, ::2, 1::2, :], inputs[:, 1::2, 1::2, :]]
+        return self.conv(tf.concat(inputs, 3))
+
+
+class TFBottleneck(keras.layers.Layer):
+    # Standard bottleneck
+    def __init__(self, c1, c2, shortcut=True, g=1, e=0.5, w=None):  # ch_in, ch_out, shortcut, groups, expansion
+        super().__init__()
+        c_ = int(c2 * e)  # hidden channels
+        self.cv1 = TFConv(c1, c_, 1, 1, w=w.cv1)
+        self.cv2 = TFConv(c_, c2, 3, 1, g=g, w=w.cv2)
+        self.add = shortcut and c1 == c2
+
+    def call(self, inputs):
+        return inputs + self.cv2(self.cv1(inputs)) if self.add else self.cv2(self.cv1(inputs))
+
+
+class TFCrossConv(keras.layers.Layer):
+    # Cross Convolution
+    def __init__(self, c1, c2, k=3, s=1, g=1, e=1.0, shortcut=False, w=None):
+        super().__init__()
+        c_ = int(c2 * e)  # hidden channels
+        self.cv1 = TFConv(c1, c_, (1, k), (1, s), w=w.cv1)
+        self.cv2 = TFConv(c_, c2, (k, 1), (s, 1), g=g, w=w.cv2)
+        self.add = shortcut and c1 == c2
+
+    def call(self, inputs):
+        return inputs + self.cv2(self.cv1(inputs)) if self.add else self.cv2(self.cv1(inputs))
+
+
+class TFConv2d(keras.layers.Layer):
+    # Substitution for PyTorch nn.Conv2D
+    def __init__(self, c1, c2, k, s=1, g=1, bias=True, w=None):
+        super().__init__()
+        assert g == 1, "TF v2.2 Conv2D does not support 'groups' argument"
+        self.conv = keras.layers.Conv2D(filters=c2,
+                                        kernel_size=k,
+                                        strides=s,
+                                        padding='VALID',
+                                        use_bias=bias,
+                                        kernel_initializer=keras.initializers.Constant(
+                                            w.weight.permute(2, 3, 1, 0).numpy()),
+                                        bias_initializer=keras.initializers.Constant(w.bias.numpy()) if bias else None)
+
+    def call(self, inputs):
+        return self.conv(inputs)
+
+
+class TFBottleneckCSP(keras.layers.Layer):
+    # CSP Bottleneck https://github.com/WongKinYiu/CrossStagePartialNetworks
+    def __init__(self, c1, c2, n=1, shortcut=True, g=1, e=0.5, w=None):
+        # ch_in, ch_out, number, shortcut, groups, expansion
+        super().__init__()
+        c_ = int(c2 * e)  # hidden channels
+        self.cv1 = TFConv(c1, c_, 1, 1, w=w.cv1)
+        self.cv2 = TFConv2d(c1, c_, 1, 1, bias=False, w=w.cv2)
+        self.cv3 = TFConv2d(c_, c_, 1, 1, bias=False, w=w.cv3)
+        self.cv4 = TFConv(2 * c_, c2, 1, 1, w=w.cv4)
+        self.bn = TFBN(w.bn)
+        self.act = lambda x: keras.activations.swish(x)
+        self.m = keras.Sequential([TFBottleneck(c_, c_, shortcut, g, e=1.0, w=w.m[j]) for j in range(n)])
+
+    def call(self, inputs):
+        y1 = self.cv3(self.m(self.cv1(inputs)))
+        y2 = self.cv2(inputs)
+        return self.cv4(self.act(self.bn(tf.concat((y1, y2), axis=3))))
+
+
+class TFC3(keras.layers.Layer):
+    # CSP Bottleneck with 3 convolutions
+    def __init__(self, c1, c2, n=1, shortcut=True, g=1, e=0.5, w=None):
+        # ch_in, ch_out, number, shortcut, groups, expansion
+        super().__init__()
+        c_ = int(c2 * e)  # hidden channels
+        self.cv1 = TFConv(c1, c_, 1, 1, w=w.cv1)
+        self.cv2 = TFConv(c1, c_, 1, 1, w=w.cv2)
+        self.cv3 = TFConv(2 * c_, c2, 1, 1, w=w.cv3)
+        self.m = keras.Sequential([TFBottleneck(c_, c_, shortcut, g, e=1.0, w=w.m[j]) for j in range(n)])
+
+    def call(self, inputs):
+        return self.cv3(tf.concat((self.m(self.cv1(inputs)), self.cv2(inputs)), axis=3))
+
+
+class TFC3x(keras.layers.Layer):
+    # 3 module with cross-convolutions
+    def __init__(self, c1, c2, n=1, shortcut=True, g=1, e=0.5, w=None):
+        # ch_in, ch_out, number, shortcut, groups, expansion
+        super().__init__()
+        c_ = int(c2 * e)  # hidden channels
+        self.cv1 = TFConv(c1, c_, 1, 1, w=w.cv1)
+        self.cv2 = TFConv(c1, c_, 1, 1, w=w.cv2)
+        self.cv3 = TFConv(2 * c_, c2, 1, 1, w=w.cv3)
+        self.m = keras.Sequential([
+            TFCrossConv(c_, c_, k=3, s=1, g=g, e=1.0, shortcut=shortcut, w=w.m[j]) for j in range(n)])
+
+    def call(self, inputs):
+        return self.cv3(tf.concat((self.m(self.cv1(inputs)), self.cv2(inputs)), axis=3))
+
+
+class TFSPP(keras.layers.Layer):
+    # Spatial pyramid pooling layer used in YOLOv3-SPP
+    def __init__(self, c1, c2, k=(5, 9, 13), w=None):
+        super().__init__()
+        c_ = c1 // 2  # hidden channels
+        self.cv1 = TFConv(c1, c_, 1, 1, w=w.cv1)
+        self.cv2 = TFConv(c_ * (len(k) + 1), c2, 1, 1, w=w.cv2)
+        self.m = [keras.layers.MaxPool2D(pool_size=x, strides=1, padding='SAME') for x in k]
+
+    def call(self, inputs):
+        x = self.cv1(inputs)
+        return self.cv2(tf.concat([x] + [m(x) for m in self.m], 3))
+
+
+class TFSPPF(keras.layers.Layer):
+    # Spatial pyramid pooling-Fast layer
+    def __init__(self, c1, c2, k=5, w=None):
+        super().__init__()
+        c_ = c1 // 2  # hidden channels
+        self.cv1 = TFConv(c1, c_, 1, 1, w=w.cv1)
+        self.cv2 = TFConv(c_ * 4, c2, 1, 1, w=w.cv2)
+        self.m = keras.layers.MaxPool2D(pool_size=k, strides=1, padding='SAME')
+
+    def call(self, inputs):
+        x = self.cv1(inputs)
+        y1 = self.m(x)
+        y2 = self.m(y1)
+        return self.cv2(tf.concat([x, y1, y2, self.m(y2)], 3))
+
+
+class TFDetect(keras.layers.Layer):
+    # TF YOLOv5 Detect layer
+    def __init__(self, nc=80, anchors=(), ch=(), imgsz=(640, 640), w=None):  # detection layer
+        super().__init__()
+        self.stride = tf.convert_to_tensor(w.stride.numpy(), dtype=tf.float32)
+        self.nc = nc  # number of classes
+        self.no = nc + 5  # number of outputs per anchor
+        self.nl = len(anchors)  # number of detection layers
+        self.na = len(anchors[0]) // 2  # number of anchors
+        self.grid = [tf.zeros(1)] * self.nl  # init grid
+        self.anchors = tf.convert_to_tensor(w.anchors.numpy(), dtype=tf.float32)
+        self.anchor_grid = tf.reshape(self.anchors * tf.reshape(self.stride, [self.nl, 1, 1]), [self.nl, 1, -1, 1, 2])
+        self.m = [TFConv2d(x, self.no * self.na, 1, w=w.m[i]) for i, x in enumerate(ch)]
+        self.training = False  # set to False after building model
+        self.imgsz = imgsz
+        for i in range(self.nl):
+            ny, nx = self.imgsz[0] // self.stride[i], self.imgsz[1] // self.stride[i]
+            self.grid[i] = self._make_grid(nx, ny)
+
+    def call(self, inputs):
+        z = []  # inference output
+        x = []
+        for i in range(self.nl):
+            x.append(self.m[i](inputs[i]))
+            # x(bs,20,20,255) to x(bs,3,20,20,85)
+            ny, nx = self.imgsz[0] // self.stride[i], self.imgsz[1] // self.stride[i]
+            x[i] = tf.reshape(x[i], [-1, ny * nx, self.na, self.no])
+
+            if not self.training:  # inference
+                y = x[i]
+                grid = tf.transpose(self.grid[i], [0, 2, 1, 3]) - 0.5
+                anchor_grid = tf.transpose(self.anchor_grid[i], [0, 2, 1, 3]) * 4
+                xy = (tf.sigmoid(y[..., 0:2]) * 2 + grid) * self.stride[i]  # xy
+                wh = tf.sigmoid(y[..., 2:4]) ** 2 * anchor_grid
+                # Normalize xywh to 0-1 to reduce calibration error
+                xy /= tf.constant([[self.imgsz[1], self.imgsz[0]]], dtype=tf.float32)
+                wh /= tf.constant([[self.imgsz[1], self.imgsz[0]]], dtype=tf.float32)
+                y = tf.concat([xy, wh, tf.sigmoid(y[..., 4:5 + self.nc]), y[..., 5 + self.nc:]], -1)
+                z.append(tf.reshape(y, [-1, self.na * ny * nx, self.no]))
+
+        return tf.transpose(x, [0, 2, 1, 3]) if self.training else (tf.concat(z, 1),)
+
+    @staticmethod
+    def _make_grid(nx=20, ny=20):
+        # yv, xv = torch.meshgrid([torch.arange(ny), torch.arange(nx)])
+        # return torch.stack((xv, yv), 2).view((1, 1, ny, nx, 2)).float()
+        xv, yv = tf.meshgrid(tf.range(nx), tf.range(ny))
+        return tf.cast(tf.reshape(tf.stack([xv, yv], 2), [1, 1, ny * nx, 2]), dtype=tf.float32)
+
+
+class TFSegment(TFDetect):
+    # YOLOv5 Segment head for segmentation models
+    def __init__(self, nc=80, anchors=(), nm=32, npr=256, ch=(), imgsz=(640, 640), w=None):
+        super().__init__(nc, anchors, ch, imgsz, w)
+        self.nm = nm  # number of masks
+        self.npr = npr  # number of protos
+        self.no = 5 + nc + self.nm  # number of outputs per anchor
+        self.m = [TFConv2d(x, self.no * self.na, 1, w=w.m[i]) for i, x in enumerate(ch)]  # output conv
+        self.proto = TFProto(ch[0], self.npr, self.nm, w=w.proto)  # protos
+        self.detect = TFDetect.call
+
+    def call(self, x):
+        p = self.proto(x[0])
+        # p = TFUpsample(None, scale_factor=4, mode='nearest')(self.proto(x[0]))  # (optional) full-size protos
+        p = tf.transpose(p, [0, 3, 1, 2])  # from shape(1,160,160,32) to shape(1,32,160,160)
+        x = self.detect(self, x)
+        return (x, p) if self.training else (x[0], p)
+
+
+class TFProto(keras.layers.Layer):
+
+    def __init__(self, c1, c_=256, c2=32, w=None):
+        super().__init__()
+        self.cv1 = TFConv(c1, c_, k=3, w=w.cv1)
+        self.upsample = TFUpsample(None, scale_factor=2, mode='nearest')
+        self.cv2 = TFConv(c_, c_, k=3, w=w.cv2)
+        self.cv3 = TFConv(c_, c2, w=w.cv3)
+
+    def call(self, inputs):
+        return self.cv3(self.cv2(self.upsample(self.cv1(inputs))))
+
+
+class TFUpsample(keras.layers.Layer):
+    # TF version of torch.nn.Upsample()
+    def __init__(self, size, scale_factor, mode, w=None):  # warning: all arguments needed including 'w'
+        super().__init__()
+        assert scale_factor % 2 == 0, "scale_factor must be multiple of 2"
+        self.upsample = lambda x: tf.image.resize(x, (x.shape[1] * scale_factor, x.shape[2] * scale_factor), mode)
+        # self.upsample = keras.layers.UpSampling2D(size=scale_factor, interpolation=mode)
+        # with default arguments: align_corners=False, half_pixel_centers=False
+        # self.upsample = lambda x: tf.raw_ops.ResizeNearestNeighbor(images=x,
+        #                                                            size=(x.shape[1] * 2, x.shape[2] * 2))
+
+    def call(self, inputs):
+        return self.upsample(inputs)
+
+
+class TFConcat(keras.layers.Layer):
+    # TF version of torch.concat()
+    def __init__(self, dimension=1, w=None):
+        super().__init__()
+        assert dimension == 1, "convert only NCHW to NHWC concat"
+        self.d = 3
+
+    def call(self, inputs):
+        return tf.concat(inputs, self.d)
+
+
+def parse_model(d, ch, model, imgsz):  # model_dict, input_channels(3)
+    LOGGER.info(f"\n{'':>3}{'from':>18}{'n':>3}{'params':>10}  {'module':<40}{'arguments':<30}")
+    anchors, nc, gd, gw = d['anchors'], d['nc'], d['depth_multiple'], d['width_multiple']
+    na = (len(anchors[0]) // 2) if isinstance(anchors, list) else anchors  # number of anchors
+    no = na * (nc + 5)  # number of outputs = anchors * (classes + 5)
+
+    layers, save, c2 = [], [], ch[-1]  # layers, savelist, ch out
+    for i, (f, n, m, args) in enumerate(d['backbone'] + d['head']):  # from, number, module, args
+        m_str = m
+        m = eval(m) if isinstance(m, str) else m  # eval strings
+        for j, a in enumerate(args):
+            try:
+                args[j] = eval(a) if isinstance(a, str) else a  # eval strings
+            except NameError:
+                pass
+
+        n = max(round(n * gd), 1) if n > 1 else n  # depth gain
+        if m in [
+                nn.Conv2d, Conv, DWConv, DWConvTranspose2d, Bottleneck, SPP, SPPF, MixConv2d, Focus, CrossConv,
+                BottleneckCSP, C3, C3x]:
+            c1, c2 = ch[f], args[0]
+            c2 = make_divisible(c2 * gw, 8) if c2 != no else c2
+
+            args = [c1, c2, *args[1:]]
+            if m in [BottleneckCSP, C3, C3x]:
+                args.insert(2, n)
+                n = 1
+        elif m is nn.BatchNorm2d:
+            args = [ch[f]]
+        elif m is Concat:
+            c2 = sum(ch[-1 if x == -1 else x + 1] for x in f)
+        elif m in [Detect, Segment]:
+            args.append([ch[x + 1] for x in f])
+            if isinstance(args[1], int):  # number of anchors
+                args[1] = [list(range(args[1] * 2))] * len(f)
+            if m is Segment:
+                args[3] = make_divisible(args[3] * gw, 8)
+            args.append(imgsz)
+        else:
+            c2 = ch[f]
+
+        tf_m = eval('TF' + m_str.replace('nn.', ''))
+        m_ = keras.Sequential([tf_m(*args, w=model.model[i][j]) for j in range(n)]) if n > 1 \
+            else tf_m(*args, w=model.model[i])  # module
+
+        torch_m_ = nn.Sequential(*(m(*args) for _ in range(n))) if n > 1 else m(*args)  # module
+        t = str(m)[8:-2].replace('__main__.', '')  # module type
+        np = sum(x.numel() for x in torch_m_.parameters())  # number params
+        m_.i, m_.f, m_.type, m_.np = i, f, t, np  # attach index, 'from' index, type, number params
+        LOGGER.info(f'{i:>3}{str(f):>18}{str(n):>3}{np:>10}  {t:<40}{str(args):<30}')  # print
+        save.extend(x % i for x in ([f] if isinstance(f, int) else f) if x != -1)  # append to savelist
+        layers.append(m_)
+        ch.append(c2)
+    return keras.Sequential(layers), sorted(save)
+
+
+class TFModel:
+    # TF YOLOv5 model
+    def __init__(self, cfg='yolov5s.yaml', ch=3, nc=None, model=None, imgsz=(640, 640)):  # model, channels, classes
+        super().__init__()
+        if isinstance(cfg, dict):
+            self.yaml = cfg  # model dict
+        else:  # is *.yaml
+            import yaml  # for torch hub
+            self.yaml_file = Path(cfg).name
+            with open(cfg) as f:
+                self.yaml = yaml.load(f, Loader=yaml.FullLoader)  # model dict
+
+        # Define model
+        if nc and nc != self.yaml['nc']:
+            LOGGER.info(f"Overriding {cfg} nc={self.yaml['nc']} with nc={nc}")
+            self.yaml['nc'] = nc  # override yaml value
+        self.model, self.savelist = parse_model(deepcopy(self.yaml), ch=[ch], model=model, imgsz=imgsz)
+
+    def predict(self,
+                inputs,
+                tf_nms=False,
+                agnostic_nms=False,
+                topk_per_class=100,
+                topk_all=100,
+                iou_thres=0.45,
+                conf_thres=0.25):
+        y = []  # outputs
+        x = inputs
+        for m in self.model.layers:
+            if m.f != -1:  # if not from previous layer
+                x = y[m.f] if isinstance(m.f, int) else [x if j == -1 else y[j] for j in m.f]  # from earlier layers
+
+            x = m(x)  # run
+            y.append(x if m.i in self.savelist else None)  # save output
+
+        # Add TensorFlow NMS
+        if tf_nms:
+            boxes = self._xywh2xyxy(x[0][..., :4])
+            probs = x[0][:, :, 4:5]
+            classes = x[0][:, :, 5:]
+            scores = probs * classes
+            if agnostic_nms:
+                nms = AgnosticNMS()((boxes, classes, scores), topk_all, iou_thres, conf_thres)
+            else:
+                boxes = tf.expand_dims(boxes, 2)
+                nms = tf.image.combined_non_max_suppression(boxes,
+                                                            scores,
+                                                            topk_per_class,
+                                                            topk_all,
+                                                            iou_thres,
+                                                            conf_thres,
+                                                            clip_boxes=False)
+            return (nms,)
+        return x  # output [1,6300,85] = [xywh, conf, class0, class1, ...]
+        # x = x[0]  # [x(1,6300,85), ...] to x(6300,85)
+        # xywh = x[..., :4]  # x(6300,4) boxes
+        # conf = x[..., 4:5]  # x(6300,1) confidences
+        # cls = tf.reshape(tf.cast(tf.argmax(x[..., 5:], axis=1), tf.float32), (-1, 1))  # x(6300,1)  classes
+        # return tf.concat([conf, cls, xywh], 1)
+
+    @staticmethod
+    def _xywh2xyxy(xywh):
+        # Convert nx4 boxes from [x, y, w, h] to [x1, y1, x2, y2] where xy1=top-left, xy2=bottom-right
+        x, y, w, h = tf.split(xywh, num_or_size_splits=4, axis=-1)
+        return tf.concat([x - w / 2, y - h / 2, x + w / 2, y + h / 2], axis=-1)
+
+
+class AgnosticNMS(keras.layers.Layer):
+    # TF Agnostic NMS
+    def call(self, input, topk_all, iou_thres, conf_thres):
+        # wrap map_fn to avoid TypeSpec related error https://stackoverflow.com/a/65809989/3036450
+        return tf.map_fn(lambda x: self._nms(x, topk_all, iou_thres, conf_thres),
+                         input,
+                         fn_output_signature=(tf.float32, tf.float32, tf.float32, tf.int32),
+                         name='agnostic_nms')
+
+    @staticmethod
+    def _nms(x, topk_all=100, iou_thres=0.45, conf_thres=0.25):  # agnostic NMS
+        boxes, classes, scores = x
+        class_inds = tf.cast(tf.argmax(classes, axis=-1), tf.float32)
+        scores_inp = tf.reduce_max(scores, -1)
+        selected_inds = tf.image.non_max_suppression(boxes,
+                                                     scores_inp,
+                                                     max_output_size=topk_all,
+                                                     iou_threshold=iou_thres,
+                                                     score_threshold=conf_thres)
+        selected_boxes = tf.gather(boxes, selected_inds)
+        padded_boxes = tf.pad(selected_boxes,
+                              paddings=[[0, topk_all - tf.shape(selected_boxes)[0]], [0, 0]],
+                              mode="CONSTANT",
+                              constant_values=0.0)
+        selected_scores = tf.gather(scores_inp, selected_inds)
+        padded_scores = tf.pad(selected_scores,
+                               paddings=[[0, topk_all - tf.shape(selected_boxes)[0]]],
+                               mode="CONSTANT",
+                               constant_values=-1.0)
+        selected_classes = tf.gather(class_inds, selected_inds)
+        padded_classes = tf.pad(selected_classes,
+                                paddings=[[0, topk_all - tf.shape(selected_boxes)[0]]],
+                                mode="CONSTANT",
+                                constant_values=-1.0)
+        valid_detections = tf.shape(selected_inds)[0]
+        return padded_boxes, padded_scores, padded_classes, valid_detections
+
+
+def activations(act=nn.SiLU):
+    # Returns TF activation from input PyTorch activation
+    if isinstance(act, nn.LeakyReLU):
+        return lambda x: keras.activations.relu(x, alpha=0.1)
+    elif isinstance(act, nn.Hardswish):
+        return lambda x: x * tf.nn.relu6(x + 3) * 0.166666667
+    elif isinstance(act, (nn.SiLU, SiLU)):
+        return lambda x: keras.activations.swish(x)
+    else:
+        raise Exception(f'no matching TensorFlow activation found for PyTorch activation {act}')
+
+
+def representative_dataset_gen(dataset, ncalib=100):
+    # Representative dataset generator for use with converter.representative_dataset, returns a generator of np arrays
+    for n, (path, img, im0s, vid_cap, string) in enumerate(dataset):
+        im = np.transpose(img, [1, 2, 0])
+        im = np.expand_dims(im, axis=0).astype(np.float32)
+        im /= 255
+        yield [im]
+        if n >= ncalib:
+            break
+
+
+def run(
+        weights=ROOT / 'yolov5s.pt',  # weights path
+        imgsz=(640, 640),  # inference size h,w
+        batch_size=1,  # batch size
+        dynamic=False,  # dynamic batch size
+):
+    # PyTorch model
+    im = torch.zeros((batch_size, 3, *imgsz))  # BCHW image
+    model = attempt_load(weights, device=torch.device('cpu'), inplace=True, fuse=False)
+    _ = model(im)  # inference
+    model.info()
+
+    # TensorFlow model
+    im = tf.zeros((batch_size, *imgsz, 3))  # BHWC image
+    tf_model = TFModel(cfg=model.yaml, model=model, nc=model.nc, imgsz=imgsz)
+    _ = tf_model.predict(im)  # inference
+
+    # Keras model
+    im = keras.Input(shape=(*imgsz, 3), batch_size=None if dynamic else batch_size)
+    keras_model = keras.Model(inputs=im, outputs=tf_model.predict(im))
+    keras_model.summary()
+
+    LOGGER.info('PyTorch, TensorFlow and Keras models successfully verified.\nUse export.py for TF model export.')
+
+
+def parse_opt():
+    parser = argparse.ArgumentParser()
+    parser.add_argument('--weights', type=str, default=ROOT / 'yolov5s.pt', help='weights path')
+    parser.add_argument('--imgsz', '--img', '--img-size', nargs='+', type=int, default=[640], help='inference size h,w')
+    parser.add_argument('--batch-size', type=int, default=1, help='batch size')
+    parser.add_argument('--dynamic', action='store_true', help='dynamic batch size')
+    opt = parser.parse_args()
+    opt.imgsz *= 2 if len(opt.imgsz) == 1 else 1  # expand
+    print_args(vars(opt))
+    return opt
+
+
+def main(opt):
+    run(**vars(opt))
+
+
+if __name__ == "__main__":
+    opt = parse_opt()
+    main(opt)

models/yolo.py

+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+"""
+YOLO-specific modules
+
+Usage:
+    $ python models/yolo.py --cfg yolov5s.yaml
+"""
+
+import argparse
+import contextlib
+import os
+import platform
+import sys
+from copy import deepcopy
+from pathlib import Path
+
+FILE = Path(__file__).resolve()
+ROOT = FILE.parents[1]  # YOLOv5 root directory
+if str(ROOT) not in sys.path:
+    sys.path.append(str(ROOT))  # add ROOT to PATH
+if platform.system() != 'Windows':
+    ROOT = Path(os.path.relpath(ROOT, Path.cwd()))  # relative
+
+from models.common import *
+from models.experimental import *
+from utils.autoanchor import check_anchor_order
+from utils.general import LOGGER, check_version, check_yaml, make_divisible, print_args
+from utils.plots import feature_visualization
+from utils.torch_utils import (fuse_conv_and_bn, initialize_weights, model_info, profile, scale_img, select_device,
+                               time_sync)
+
+try:
+    import thop  # for FLOPs computation
+except ImportError:
+    thop = None
+
+
+class Detect(nn.Module):
+    # YOLOv5 Detect head for detection models
+    stride = None  # strides computed during build
+    dynamic = False  # force grid reconstruction
+    export = False  # export mode
+
+    def __init__(self, nc=80, anchors=(), ch=(), inplace=True):  # detection layer
+        super().__init__()
+        self.nc = nc  # number of classes
+        self.no = nc + 5  # number of outputs per anchor
+        self.nl = len(anchors)  # number of detection layers
+        self.na = len(anchors[0]) // 2  # number of anchors
+        self.grid = [torch.empty(0) for _ in range(self.nl)]  # init grid
+        self.anchor_grid = [torch.empty(0) for _ in range(self.nl)]  # init anchor grid
+        self.register_buffer('anchors', torch.tensor(anchors).float().view(self.nl, -1, 2))  # shape(nl,na,2)
+        self.m = nn.ModuleList(nn.Conv2d(x, self.no * self.na, 1) for x in ch)  # output conv
+        self.inplace = inplace  # use inplace ops (e.g. slice assignment)
+
+    def forward(self, x):
+        z = []  # inference output
+        for i in range(self.nl):
+            x[i] = self.m[i](x[i])  # conv
+            bs, _, ny, nx = x[i].shape  # x(bs,255,20,20) to x(bs,3,20,20,85)
+            x[i] = x[i].view(bs, self.na, self.no, ny, nx).permute(0, 1, 3, 4, 2).contiguous()
+
+            if not self.training:  # inference
+                if self.dynamic or self.grid[i].shape[2:4] != x[i].shape[2:4]:
+                    self.grid[i], self.anchor_grid[i] = self._make_grid(nx, ny, i)
+
+                if isinstance(self, Segment):  # (boxes + masks)
+                    xy, wh, conf, mask = x[i].split((2, 2, self.nc + 1, self.no - self.nc - 5), 4)
+                    xy = (xy.sigmoid() * 2 + self.grid[i]) * self.stride[i]  # xy
+                    wh = (wh.sigmoid() * 2) ** 2 * self.anchor_grid[i]  # wh
+                    y = torch.cat((xy, wh, conf.sigmoid(), mask), 4)
+                else:  # Detect (boxes only)
+                    xy, wh, conf = x[i].sigmoid().split((2, 2, self.nc + 1), 4)
+                    xy = (xy * 2 + self.grid[i]) * self.stride[i]  # xy
+                    wh = (wh * 2) ** 2 * self.anchor_grid[i]  # wh
+                    y = torch.cat((xy, wh, conf), 4)
+                z.append(y.view(bs, self.na * nx * ny, self.no))
+
+        return x if self.training else (torch.cat(z, 1),) if self.export else (torch.cat(z, 1), x)
+
+    def _make_grid(self, nx=20, ny=20, i=0, torch_1_10=check_version(torch.__version__, '1.10.0')):
+        d = self.anchors[i].device
+        t = self.anchors[i].dtype
+        shape = 1, self.na, ny, nx, 2  # grid shape
+        y, x = torch.arange(ny, device=d, dtype=t), torch.arange(nx, device=d, dtype=t)
+        yv, xv = torch.meshgrid(y, x, indexing='ij') if torch_1_10 else torch.meshgrid(y, x)  # torch>=0.7 compatibility
+        grid = torch.stack((xv, yv), 2).expand(shape) - 0.5  # add grid offset, i.e. y = 2.0 * x - 0.5
+        anchor_grid = (self.anchors[i] * self.stride[i]).view((1, self.na, 1, 1, 2)).expand(shape)
+        return grid, anchor_grid
+
+
+class Segment(Detect):
+    # YOLOv5 Segment head for segmentation models
+    def __init__(self, nc=80, anchors=(), nm=32, npr=256, ch=(), inplace=True):
+        super().__init__(nc, anchors, ch, inplace)
+        self.nm = nm  # number of masks
+        self.npr = npr  # number of protos
+        self.no = 5 + nc + self.nm  # number of outputs per anchor
+        self.m = nn.ModuleList(nn.Conv2d(x, self.no * self.na, 1) for x in ch)  # output conv
+        self.proto = Proto(ch[0], self.npr, self.nm)  # protos
+        self.detect = Detect.forward
+
+    def forward(self, x):
+        p = self.proto(x[0])
+        x = self.detect(self, x)
+        return (x, p) if self.training else (x[0], p) if self.export else (x[0], p, x[1])
+
+
+class BaseModel(nn.Module):
+    # YOLOv5 base model
+    def forward(self, x, profile=False, visualize=False):
+        return self._forward_once(x, profile, visualize)  # single-scale inference, train
+
+    def _forward_once(self, x, profile=False, visualize=False):
+        y, dt = [], []  # outputs
+        for m in self.model:
+            if m.f != -1:  # if not from previous layer
+                x = y[m.f] if isinstance(m.f, int) else [x if j == -1 else y[j] for j in m.f]  # from earlier layers
+            if profile:
+                self._profile_one_layer(m, x, dt)
+            x = m(x)  # run
+            y.append(x if m.i in self.save else None)  # save output
+            if visualize:
+                feature_visualization(x, m.type, m.i, save_dir=visualize)
+        return x
+
+    def _profile_one_layer(self, m, x, dt):
+        c = m == self.model[-1]  # is final layer, copy input as inplace fix
+        o = thop.profile(m, inputs=(x.copy() if c else x,), verbose=False)[0] / 1E9 * 2 if thop else 0  # FLOPs
+        t = time_sync()
+        for _ in range(10):
+            m(x.copy() if c else x)
+        dt.append((time_sync() - t) * 100)
+        if m == self.model[0]:
+            LOGGER.info(f"{'time (ms)':>10s} {'GFLOPs':>10s} {'params':>10s}  module")
+        LOGGER.info(f'{dt[-1]:10.2f} {o:10.2f} {m.np:10.0f}  {m.type}')
+        if c:
+            LOGGER.info(f"{sum(dt):10.2f} {'-':>10s} {'-':>10s}  Total")
+
+    def fuse(self):  # fuse model Conv2d() + BatchNorm2d() layers
+        LOGGER.info('Fusing layers... ')
+        for m in self.model.modules():
+            if isinstance(m, (Conv, DWConv)) and hasattr(m, 'bn'):
+                m.conv = fuse_conv_and_bn(m.conv, m.bn)  # update conv
+                delattr(m, 'bn')  # remove batchnorm
+                m.forward = m.forward_fuse  # update forward
+        self.info()
+        return self
+
+    def info(self, verbose=False, img_size=640):  # print model information
+        model_info(self, verbose, img_size)
+
+    def _apply(self, fn):
+        # Apply to(), cpu(), cuda(), half() to model tensors that are not parameters or registered buffers
+        self = super()._apply(fn)
+        m = self.model[-1]  # Detect()
+        if isinstance(m, (Detect, Segment)):
+            m.stride = fn(m.stride)
+            m.grid = list(map(fn, m.grid))
+            if isinstance(m.anchor_grid, list):
+                m.anchor_grid = list(map(fn, m.anchor_grid))
+        return self
+
+
+class DetectionModel(BaseModel):
+    # YOLOv5 detection model
+    def __init__(self, cfg='yolov5s.yaml', ch=3, nc=None, anchors=None):  # model, input channels, number of classes
+        super().__init__()
+        if isinstance(cfg, dict):
+            self.yaml = cfg  # model dict
+        else:  # is *.yaml
+            import yaml  # for torch hub
+            self.yaml_file = Path(cfg).name
+            with open(cfg, encoding='ascii', errors='ignore') as f:
+                self.yaml = yaml.safe_load(f)  # model dict
+
+        # Define model
+        ch = self.yaml['ch'] = self.yaml.get('ch', ch)  # input channels
+        if nc and nc != self.yaml['nc']:
+            LOGGER.info(f"Overriding model.yaml nc={self.yaml['nc']} with nc={nc}")
+            self.yaml['nc'] = nc  # override yaml value
+        if anchors:
+            LOGGER.info(f'Overriding model.yaml anchors with anchors={anchors}')
+            self.yaml['anchors'] = round(anchors)  # override yaml value
+        self.model, self.save = parse_model(deepcopy(self.yaml), ch=[ch])  # model, savelist
+        self.names = [str(i) for i in range(self.yaml['nc'])]  # default names
+        self.inplace = self.yaml.get('inplace', True)
+
+        # Build strides, anchors
+        m = self.model[-1]  # Detect()
+        if isinstance(m, (Detect, Segment)):
+            s = 256  # 2x min stride
+            m.inplace = self.inplace
+            forward = lambda x: self.forward(x)[0] if isinstance(m, Segment) else self.forward(x)
+            m.stride = torch.tensor([s / x.shape[-2] for x in forward(torch.zeros(1, ch, s, s))])  # forward
+            check_anchor_order(m)
+            m.anchors /= m.stride.view(-1, 1, 1)
+            self.stride = m.stride
+            self._initialize_biases()  # only run once
+
+        # Init weights, biases
+        initialize_weights(self)
+        self.info()
+        LOGGER.info('')
+
+    def forward(self, x, augment=False, profile=False, visualize=False):
+        if augment:
+            return self._forward_augment(x)  # augmented inference, None
+        return self._forward_once(x, profile, visualize)  # single-scale inference, train
+
+    def _forward_augment(self, x):
+        img_size = x.shape[-2:]  # height, width
+        s = [1, 0.83, 0.67]  # scales
+        f = [None, 3, None]  # flips (2-ud, 3-lr)
+        y = []  # outputs
+        for si, fi in zip(s, f):
+            xi = scale_img(x.flip(fi) if fi else x, si, gs=int(self.stride.max()))
+            yi = self._forward_once(xi)[0]  # forward
+            # cv2.imwrite(f'img_{si}.jpg', 255 * xi[0].cpu().numpy().transpose((1, 2, 0))[:, :, ::-1])  # save
+            yi = self._descale_pred(yi, fi, si, img_size)
+            y.append(yi)
+        y = self._clip_augmented(y)  # clip augmented tails
+        return torch.cat(y, 1), None  # augmented inference, train
+
+    def _descale_pred(self, p, flips, scale, img_size):
+        # de-scale predictions following augmented inference (inverse operation)
+        if self.inplace:
+            p[..., :4] /= scale  # de-scale
+            if flips == 2:
+                p[..., 1] = img_size[0] - p[..., 1]  # de-flip ud
+            elif flips == 3:
+                p[..., 0] = img_size[1] - p[..., 0]  # de-flip lr
+        else:
+            x, y, wh = p[..., 0:1] / scale, p[..., 1:2] / scale, p[..., 2:4] / scale  # de-scale
+            if flips == 2:
+                y = img_size[0] - y  # de-flip ud
+            elif flips == 3:
+                x = img_size[1] - x  # de-flip lr
+            p = torch.cat((x, y, wh, p[..., 4:]), -1)
+        return p
+
+    def _clip_augmented(self, y):
+        # Clip YOLOv5 augmented inference tails
+        nl = self.model[-1].nl  # number of detection layers (P3-P5)
+        g = sum(4 ** x for x in range(nl))  # grid points
+        e = 1  # exclude layer count
+        i = (y[0].shape[1] // g) * sum(4 ** x for x in range(e))  # indices
+        y[0] = y[0][:, :-i]  # large
+        i = (y[-1].shape[1] // g) * sum(4 ** (nl - 1 - x) for x in range(e))  # indices
+        y[-1] = y[-1][:, i:]  # small
+        return y
+
+    def _initialize_biases(self, cf=None):  # initialize biases into Detect(), cf is class frequency
+        # https://arxiv.org/abs/1708.02002 section 3.3
+        # cf = torch.bincount(torch.tensor(np.concatenate(dataset.labels, 0)[:, 0]).long(), minlength=nc) + 1.
+        m = self.model[-1]  # Detect() module
+        for mi, s in zip(m.m, m.stride):  # from
+            b = mi.bias.view(m.na, -1)  # conv.bias(255) to (3,85)
+            b.data[:, 4] += math.log(8 / (640 / s) ** 2)  # obj (8 objects per 640 image)
+            b.data[:, 5:5 + m.nc] += math.log(0.6 / (m.nc - 0.99999)) if cf is None else torch.log(cf / cf.sum())  # cls
+            mi.bias = torch.nn.Parameter(b.view(-1), requires_grad=True)
+
+
+Model = DetectionModel  # retain YOLOv5 'Model' class for backwards compatibility
+
+
+class SegmentationModel(DetectionModel):
+    # YOLOv5 segmentation model
+    def __init__(self, cfg='yolov5s-seg.yaml', ch=3, nc=None, anchors=None):
+        super().__init__(cfg, ch, nc, anchors)
+
+
+class ClassificationModel(BaseModel):
+    # YOLOv5 classification model
+    def __init__(self, cfg=None, model=None, nc=1000, cutoff=10):  # yaml, model, number of classes, cutoff index
+        super().__init__()
+        self._from_detection_model(model, nc, cutoff) if model is not None else self._from_yaml(cfg)
+
+    def _from_detection_model(self, model, nc=1000, cutoff=10):
+        # Create a YOLOv5 classification model from a YOLOv5 detection model
+        if isinstance(model, DetectMultiBackend):
+            model = model.model  # unwrap DetectMultiBackend
+        model.model = model.model[:cutoff]  # backbone
+        m = model.model[-1]  # last layer
+        ch = m.conv.in_channels if hasattr(m, 'conv') else m.cv1.conv.in_channels  # ch into module
+        c = Classify(ch, nc)  # Classify()
+        c.i, c.f, c.type = m.i, m.f, 'models.common.Classify'  # index, from, type
+        model.model[-1] = c  # replace
+        self.model = model.model
+        self.stride = model.stride
+        self.save = []
+        self.nc = nc
+
+    def _from_yaml(self, cfg):
+        # Create a YOLOv5 classification model from a *.yaml file
+        self.model = None
+
+
+def parse_model(d, ch):  # model_dict, input_channels(3)
+    # Parse a YOLOv5 model.yaml dictionary
+    LOGGER.info(f"\n{'':>3}{'from':>18}{'n':>3}{'params':>10}  {'module':<40}{'arguments':<30}")
+    anchors, nc, gd, gw, act = d['anchors'], d['nc'], d['depth_multiple'], d['width_multiple'], d.get('activation')
+    if act:
+        Conv.default_act = eval(act)  # redefine default activation, i.e. Conv.default_act = nn.SiLU()
+        LOGGER.info(f"{colorstr('activation:')} {act}")  # print
+    na = (len(anchors[0]) // 2) if isinstance(anchors, list) else anchors  # number of anchors
+    no = na * (nc + 5)  # number of outputs = anchors * (classes + 5)
+
+    layers, save, c2 = [], [], ch[-1]  # layers, savelist, ch out
+    for i, (f, n, m, args) in enumerate(d['backbone'] + d['head']):  # from, number, module, args
+        m = eval(m) if isinstance(m, str) else m  # eval strings
+        for j, a in enumerate(args):
+            with contextlib.suppress(NameError):
+                args[j] = eval(a) if isinstance(a, str) else a  # eval strings
+
+        n = n_ = max(round(n * gd), 1) if n > 1 else n  # depth gain
+        if m in {
+                Conv, GhostConv, Bottleneck, GhostBottleneck, SPP, SPPF, DWConv, MixConv2d, Focus, CrossConv,
+                BottleneckCSP, C3, C3TR, C3SPP, C3Ghost, nn.ConvTranspose2d, DWConvTranspose2d, C3x}:
+            c1, c2 = ch[f], args[0]
+            if c2 != no:  # if not output
+                c2 = make_divisible(c2 * gw, 8)
+
+            args = [c1, c2, *args[1:]]
+            if m in {BottleneckCSP, C3, C3TR, C3Ghost, C3x}:
+                args.insert(2, n)  # number of repeats
+                n = 1
+        elif m is nn.BatchNorm2d:
+            args = [ch[f]]
+        elif m is Concat:
+            c2 = sum(ch[x] for x in f)
+        # TODO: channel, gw, gd
+        elif m in {Detect, Segment}:
+            args.append([ch[x] for x in f])
+            if isinstance(args[1], int):  # number of anchors
+                args[1] = [list(range(args[1] * 2))] * len(f)
+            if m is Segment:
+                args[3] = make_divisible(args[3] * gw, 8)
+        elif m is Contract:
+            c2 = ch[f] * args[0] ** 2
+        elif m is Expand:
+            c2 = ch[f] // args[0] ** 2
+        else:
+            c2 = ch[f]
+
+        m_ = nn.Sequential(*(m(*args) for _ in range(n))) if n > 1 else m(*args)  # module
+        t = str(m)[8:-2].replace('__main__.', '')  # module type
+        np = sum(x.numel() for x in m_.parameters())  # number params
+        m_.i, m_.f, m_.type, m_.np = i, f, t, np  # attach index, 'from' index, type, number params
+        LOGGER.info(f'{i:>3}{str(f):>18}{n_:>3}{np:10.0f}  {t:<40}{str(args):<30}')  # print
+        save.extend(x % i for x in ([f] if isinstance(f, int) else f) if x != -1)  # append to savelist
+        layers.append(m_)
+        if i == 0:
+            ch = []
+        ch.append(c2)
+    return nn.Sequential(*layers), sorted(save)
+
+
+if __name__ == '__main__':
+    parser = argparse.ArgumentParser()
+    parser.add_argument('--cfg', type=str, default='yolov5s.yaml', help='model.yaml')
+    parser.add_argument('--batch-size', type=int, default=1, help='total batch size for all GPUs')
+    parser.add_argument('--device', default='', help='cuda device, i.e. 0 or 0,1,2,3 or cpu')
+    parser.add_argument('--profile', action='store_true', help='profile model speed')
+    parser.add_argument('--line-profile', action='store_true', help='profile model speed layer by layer')
+    parser.add_argument('--test', action='store_true', help='test all yolo*.yaml')
+    opt = parser.parse_args()
+    opt.cfg = check_yaml(opt.cfg)  # check YAML
+    print_args(vars(opt))
+    device = select_device(opt.device)
+
+    # Create model
+    im = torch.rand(opt.batch_size, 3, 640, 640).to(device)
+    model = Model(opt.cfg).to(device)
+
+    # Options
+    if opt.line_profile:  # profile layer by layer
+        model(im, profile=True)
+
+    elif opt.profile:  # profile forward-backward
+        results = profile(input=im, ops=[model], n=3)
+
+    elif opt.test:  # test all models
+        for cfg in Path(ROOT / 'models').rglob('yolo*.yaml'):
+            try:
+                _ = Model(cfg)
+            except Exception as e:
+                print(f'Error in {cfg}: {e}')
+
+    else:  # report fused model summary
+        model.fuse()

models/yolov5l.yaml

+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+
+# Parameters
+nc: 80  # number of classes
+depth_multiple: 1.0  # model depth multiple
+width_multiple: 1.0  # layer channel multiple
+anchors:
+  - [10,13, 16,30, 33,23]  # P3/8
+  - [30,61, 62,45, 59,119]  # P4/16
+  - [116,90, 156,198, 373,326]  # P5/32
+
+# YOLOv5 v6.0 backbone
+backbone:
+  # [from, number, module, args]
+  [[-1, 1, Conv, [64, 6, 2, 2]],  # 0-P1/2
+   [-1, 1, Conv, [128, 3, 2]],  # 1-P2/4
+   [-1, 3, C3, [128]],
+   [-1, 1, Conv, [256, 3, 2]],  # 3-P3/8
+   [-1, 6, C3, [256]],
+   [-1, 1, Conv, [512, 3, 2]],  # 5-P4/16
+   [-1, 9, C3, [512]],
+   [-1, 1, Conv, [1024, 3, 2]],  # 7-P5/32
+   [-1, 3, C3, [1024]],
+   [-1, 1, SPPF, [1024, 5]],  # 9
+  ]
+
+# YOLOv5 v6.0 head
+head:
+  [[-1, 1, Conv, [512, 1, 1]],
+   [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+   [[-1, 6], 1, Concat, [1]],  # cat backbone P4
+   [-1, 3, C3, [512, False]],  # 13
+
+   [-1, 1, Conv, [256, 1, 1]],
+   [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+   [[-1, 4], 1, Concat, [1]],  # cat backbone P3
+   [-1, 3, C3, [256, False]],  # 17 (P3/8-small)
+
+   [-1, 1, Conv, [256, 3, 2]],
+   [[-1, 14], 1, Concat, [1]],  # cat head P4
+   [-1, 3, C3, [512, False]],  # 20 (P4/16-medium)
+
+   [-1, 1, Conv, [512, 3, 2]],
+   [[-1, 10], 1, Concat, [1]],  # cat head P5
+   [-1, 3, C3, [1024, False]],  # 23 (P5/32-large)
+
+   [[17, 20, 23], 1, Detect, [nc, anchors]],  # Detect(P3, P4, P5)
+  ]

models/yolov5m.yaml

+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+
+# Parameters
+nc: 80  # number of classes
+depth_multiple: 0.67  # model depth multiple
+width_multiple: 0.75  # layer channel multiple
+anchors:
+  - [10,13, 16,30, 33,23]  # P3/8
+  - [30,61, 62,45, 59,119]  # P4/16
+  - [116,90, 156,198, 373,326]  # P5/32
+
+# YOLOv5 v6.0 backbone
+backbone:
+  # [from, number, module, args]
+  [[-1, 1, Conv, [64, 6, 2, 2]],  # 0-P1/2
+   [-1, 1, Conv, [128, 3, 2]],  # 1-P2/4
+   [-1, 3, C3, [128]],
+   [-1, 1, Conv, [256, 3, 2]],  # 3-P3/8
+   [-1, 6, C3, [256]],
+   [-1, 1, Conv, [512, 3, 2]],  # 5-P4/16
+   [-1, 9, C3, [512]],
+   [-1, 1, Conv, [1024, 3, 2]],  # 7-P5/32
+   [-1, 3, C3, [1024]],
+   [-1, 1, SPPF, [1024, 5]],  # 9
+  ]
+
+# YOLOv5 v6.0 head
+head:
+  [[-1, 1, Conv, [512, 1, 1]],
+   [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+   [[-1, 6], 1, Concat, [1]],  # cat backbone P4
+   [-1, 3, C3, [512, False]],  # 13
+
+   [-1, 1, Conv, [256, 1, 1]],
+   [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+   [[-1, 4], 1, Concat, [1]],  # cat backbone P3
+   [-1, 3, C3, [256, False]],  # 17 (P3/8-small)
+
+   [-1, 1, Conv, [256, 3, 2]],
+   [[-1, 14], 1, Concat, [1]],  # cat head P4
+   [-1, 3, C3, [512, False]],  # 20 (P4/16-medium)
+
+   [-1, 1, Conv, [512, 3, 2]],
+   [[-1, 10], 1, Concat, [1]],  # cat head P5
+   [-1, 3, C3, [1024, False]],  # 23 (P5/32-large)
+
+   [[17, 20, 23], 1, Detect, [nc, anchors]],  # Detect(P3, P4, P5)
+  ]

models/yolov5n.yaml

+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+
+# Parameters
+nc: 80  # number of classes
+depth_multiple: 0.33  # model depth multiple
+width_multiple: 0.25  # layer channel multiple
+anchors:
+  - [10,13, 16,30, 33,23]  # P3/8
+  - [30,61, 62,45, 59,119]  # P4/16
+  - [116,90, 156,198, 373,326]  # P5/32
+
+# YOLOv5 v6.0 backbone
+backbone:
+  # [from, number, module, args]
+  [[-1, 1, Conv, [64, 6, 2, 2]],  # 0-P1/2
+   [-1, 1, Conv, [128, 3, 2]],  # 1-P2/4
+   [-1, 3, C3, [128]],
+   [-1, 1, Conv, [256, 3, 2]],  # 3-P3/8
+   [-1, 6, C3, [256]],
+   [-1, 1, Conv, [512, 3, 2]],  # 5-P4/16
+   [-1, 9, C3, [512]],
+   [-1, 1, Conv, [1024, 3, 2]],  # 7-P5/32
+   [-1, 3, C3, [1024]],
+   [-1, 1, SPPF, [1024, 5]],  # 9
+  ]
+
+# YOLOv5 v6.0 head
+head:
+  [[-1, 1, Conv, [512, 1, 1]],
+   [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+   [[-1, 6], 1, Concat, [1]],  # cat backbone P4
+   [-1, 3, C3, [512, False]],  # 13
+
+   [-1, 1, Conv, [256, 1, 1]],
+   [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+   [[-1, 4], 1, Concat, [1]],  # cat backbone P3
+   [-1, 3, C3, [256, False]],  # 17 (P3/8-small)
+
+   [-1, 1, Conv, [256, 3, 2]],
+   [[-1, 14], 1, Concat, [1]],  # cat head P4
+   [-1, 3, C3, [512, False]],  # 20 (P4/16-medium)
+
+   [-1, 1, Conv, [512, 3, 2]],
+   [[-1, 10], 1, Concat, [1]],  # cat head P5
+   [-1, 3, C3, [1024, False]],  # 23 (P5/32-large)
+
+   [[17, 20, 23], 1, Detect, [nc, anchors]],  # Detect(P3, P4, P5)
+  ]

models/yolov5s.yaml

+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+
+# Parameters
+nc: 80  # number of classes
+depth_multiple: 0.33  # model depth multiple
+width_multiple: 0.50  # layer channel multiple
+anchors:
+  - [10,13, 16,30, 33,23]  # P3/8
+  - [30,61, 62,45, 59,119]  # P4/16
+  - [116,90, 156,198, 373,326]  # P5/32
+
+# YOLOv5 v6.0 backbone
+backbone:
+  # [from, number, module, args]
+  [[-1, 1, Conv, [64, 6, 2, 2]],  # 0-P1/2
+   [-1, 1, Conv, [128, 3, 2]],  # 1-P2/4
+   [-1, 3, C3, [128]],
+   [-1, 1, Conv, [256, 3, 2]],  # 3-P3/8
+   [-1, 6, C3, [256]],
+   [-1, 1, Conv, [512, 3, 2]],  # 5-P4/16
+   [-1, 9, C3, [512]],
+   [-1, 1, Conv, [1024, 3, 2]],  # 7-P5/32
+   [-1, 3, C3, [1024]],
+   [-1, 1, SPPF, [1024, 5]],  # 9
+  ]
+
+# YOLOv5 v6.0 head
+head:
+  [[-1, 1, Conv, [512, 1, 1]],
+   [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+   [[-1, 6], 1, Concat, [1]],  # cat backbone P4
+   [-1, 3, C3, [512, False]],  # 13
+
+   [-1, 1, Conv, [256, 1, 1]],
+   [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+   [[-1, 4], 1, Concat, [1]],  # cat backbone P3
+   [-1, 3, C3, [256, False]],  # 17 (P3/8-small)
+
+   [-1, 1, Conv, [256, 3, 2]],
+   [[-1, 14], 1, Concat, [1]],  # cat head P4
+   [-1, 3, C3, [512, False]],  # 20 (P4/16-medium)
+
+   [-1, 1, Conv, [512, 3, 2]],
+   [[-1, 10], 1, Concat, [1]],  # cat head P5
+   [-1, 3, C3, [1024, False]],  # 23 (P5/32-large)
+
+   [[17, 20, 23], 1, Detect, [nc, anchors]],  # Detect(P3, P4, P5)
+  ]

models/yolov5x.yaml

+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+
+# Parameters
+nc: 80  # number of classes
+depth_multiple: 1.33  # model depth multiple
+width_multiple: 1.25  # layer channel multiple
+anchors:
+  - [10,13, 16,30, 33,23]  # P3/8
+  - [30,61, 62,45, 59,119]  # P4/16
+  - [116,90, 156,198, 373,326]  # P5/32
+
+# YOLOv5 v6.0 backbone
+backbone:
+  # [from, number, module, args]
+  [[-1, 1, Conv, [64, 6, 2, 2]],  # 0-P1/2
+   [-1, 1, Conv, [128, 3, 2]],  # 1-P2/4
+   [-1, 3, C3, [128]],
+   [-1, 1, Conv, [256, 3, 2]],  # 3-P3/8
+   [-1, 6, C3, [256]],
+   [-1, 1, Conv, [512, 3, 2]],  # 5-P4/16
+   [-1, 9, C3, [512]],
+   [-1, 1, Conv, [1024, 3, 2]],  # 7-P5/32
+   [-1, 3, C3, [1024]],
+   [-1, 1, SPPF, [1024, 5]],  # 9
+  ]
+
+# YOLOv5 v6.0 head
+head:
+  [[-1, 1, Conv, [512, 1, 1]],
+   [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+   [[-1, 6], 1, Concat, [1]],  # cat backbone P4
+   [-1, 3, C3, [512, False]],  # 13
+
+   [-1, 1, Conv, [256, 1, 1]],
+   [-1, 1, nn.Upsample, [None, 2, 'nearest']],
+   [[-1, 4], 1, Concat, [1]],  # cat backbone P3
+   [-1, 3, C3, [256, False]],  # 17 (P3/8-small)
+
+   [-1, 1, Conv, [256, 3, 2]],
+   [[-1, 14], 1, Concat, [1]],  # cat head P4
+   [-1, 3, C3, [512, False]],  # 20 (P4/16-medium)
+
+   [-1, 1, Conv, [512, 3, 2]],
+   [[-1, 10], 1, Concat, [1]],  # cat head P5
+   [-1, 3, C3, [1024, False]],  # 23 (P5/32-large)
+
+   [[17, 20, 23], 1, Detect, [nc, anchors]],  # Detect(P3, P4, P5)
+  ]

requirements.txt

+# YOLOv5 🚀 requirements
+# Usage: pip install -r requirements.txt
+
+# Base ------------------------------------------------------------------------
+gitpython
+ipython  # interactive notebook
+matplotlib>=3.2.2
+numpy>=1.18.5
+opencv-python>=4.1.1
+Pillow>=7.1.2
+psutil  # system resources
+PyYAML>=5.3.1
+requests>=2.23.0
+scipy>=1.4.1
+thop>=0.1.1  # FLOPs computation
+torch>=1.7.0  # see https://pytorch.org/get-started/locally (recommended)
+torchvision>=0.8.1
+tqdm>=4.64.0
+# protobuf<=3.20.1  # https://github.com/ultralytics/yolov5/issues/8012
+
+# Logging ---------------------------------------------------------------------
+tensorboard>=2.4.1
+# clearml>=1.2.0
+# comet
+
+# Plotting --------------------------------------------------------------------
+pandas>=1.1.4
+seaborn>=0.11.0
+
+# Export ----------------------------------------------------------------------
+# coremltools>=6.0  # CoreML export
+# onnx>=1.9.0  # ONNX export
+# onnx-simplifier>=0.4.1  # ONNX simplifier
+# nvidia-pyindex  # TensorRT export
+# nvidia-tensorrt  # TensorRT export
+# scikit-learn<=1.1.2  # CoreML quantization
+# tensorflow>=2.4.1  # TF exports (-cpu, -aarch64, -macos)
+# tensorflowjs>=3.9.0  # TF.js export
+# openvino-dev  # OpenVINO export
+
+# Deploy ----------------------------------------------------------------------
+# tritonclient[all]~=2.24.0
+
+# Extras ----------------------------------------------------------------------
+# mss  # screenshots
+# albumentations>=1.0.3
+# pycocotools>=2.0  # COCO mAP
+# roboflow
+# ultralytics  # HUB https://hub.ultralytics.com

segment/predict.py

+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+"""
+Run YOLOv5 segmentation inference on images, videos, directories, streams, etc.
+
+Usage - sources:
+    $ python segment/predict.py --weights yolov5s-seg.pt --source 0                               # webcam
+                                                                  img.jpg                         # image
+                                                                  vid.mp4                         # video
+                                                                  screen                          # screenshot
+                                                                  path/                           # directory
+                                                                  'path/*.jpg'                    # glob
+                                                                  'https://youtu.be/Zgi9g1ksQHc'  # YouTube
+                                                                  'rtsp://example.com/media.mp4'  # RTSP, RTMP, HTTP stream
+
+Usage - formats:
+    $ python segment/predict.py --weights yolov5s-seg.pt                 # PyTorch
+                                          yolov5s-seg.torchscript        # TorchScript
+                                          yolov5s-seg.onnx               # ONNX Runtime or OpenCV DNN with --dnn
+                                          yolov5s-seg_openvino_model     # OpenVINO
+                                          yolov5s-seg.engine             # TensorRT
+                                          yolov5s-seg.mlmodel            # CoreML (macOS-only)
+                                          yolov5s-seg_saved_model        # TensorFlow SavedModel
+                                          yolov5s-seg.pb                 # TensorFlow GraphDef
+                                          yolov5s-seg.tflite             # TensorFlow Lite
+                                          yolov5s-seg_edgetpu.tflite     # TensorFlow Edge TPU
+                                          yolov5s-seg_paddle_model       # PaddlePaddle
+"""
+
+import argparse
+import os
+import platform
+import sys
+from pathlib import Path
+
+import torch
+
+FILE = Path(__file__).resolve()
+ROOT = FILE.parents[1]  # YOLOv5 root directory
+if str(ROOT) not in sys.path:
+    sys.path.append(str(ROOT))  # add ROOT to PATH
+ROOT = Path(os.path.relpath(ROOT, Path.cwd()))  # relative
+
+from models.common import DetectMultiBackend
+from utils.dataloaders import IMG_FORMATS, VID_FORMATS, LoadImages, LoadScreenshots, LoadStreams
+from utils.general import (LOGGER, Profile, check_file, check_img_size, check_imshow, check_requirements, colorstr, cv2,
+                           increment_path, non_max_suppression, print_args, scale_boxes, scale_segments,
+                           strip_optimizer, xyxy2xywh)
+from utils.plots import Annotator, colors, save_one_box
+from utils.segment.general import masks2segments, process_mask
+from utils.torch_utils import select_device, smart_inference_mode
+
+
+@smart_inference_mode()
+def run(
+    weights=ROOT / 'yolov5s-seg.pt',  # model.pt path(s)
+    source=ROOT / 'data/images',  # file/dir/URL/glob/screen/0(webcam)
+    data=ROOT / 'data/coco128.yaml',  # dataset.yaml path
+    imgsz=(640, 640),  # inference size (height, width)
+    conf_thres=0.25,  # confidence threshold
+    iou_thres=0.45,  # NMS IOU threshold
+    max_det=1000,  # maximum detections per image
+    device='',  # cuda device, i.e. 0 or 0,1,2,3 or cpu
+    view_img=False,  # show results
+    save_txt=False,  # save results to *.txt
+    save_conf=False,  # save confidences in --save-txt labels
+    save_crop=False,  # save cropped prediction boxes
+    nosave=False,  # do not save images/videos
+    classes=None,  # filter by class: --class 0, or --class 0 2 3
+    agnostic_nms=False,  # class-agnostic NMS
+    augment=False,  # augmented inference
+    visualize=False,  # visualize features
+    update=False,  # update all models
+    project=ROOT / 'runs/predict-seg',  # save results to project/name
+    name='exp',  # save results to project/name
+    exist_ok=False,  # existing project/name ok, do not increment
+    line_thickness=3,  # bounding box thickness (pixels)
+    hide_labels=False,  # hide labels
+    hide_conf=False,  # hide confidences
+    half=False,  # use FP16 half-precision inference
+    dnn=False,  # use OpenCV DNN for ONNX inference
+    vid_stride=1,  # video frame-rate stride
+    retina_masks=False,
+):
+    source = str(source)
+    save_img = not nosave and not source.endswith('.txt')  # save inference images
+    is_file = Path(source).suffix[1:] in (IMG_FORMATS + VID_FORMATS)
+    is_url = source.lower().startswith(('rtsp://', 'rtmp://', 'http://', 'https://'))
+    webcam = source.isnumeric() or source.endswith('.txt') or (is_url and not is_file)
+    screenshot = source.lower().startswith('screen')
+    if is_url and is_file:
+        source = check_file(source)  # download
+
+    # Directories
+    save_dir = increment_path(Path(project) / name, exist_ok=exist_ok)  # increment run
+    (save_dir / 'labels' if save_txt else save_dir).mkdir(parents=True, exist_ok=True)  # make dir
+
+    # Load model
+    device = select_device(device)
+    model = DetectMultiBackend(weights, device=device, dnn=dnn, data=data, fp16=half)
+    stride, names, pt = model.stride, model.names, model.pt
+    imgsz = check_img_size(imgsz, s=stride)  # check image size
+
+    # Dataloader
+    bs = 1  # batch_size
+    if webcam:
+        view_img = check_imshow(warn=True)
+        dataset = LoadStreams(source, img_size=imgsz, stride=stride, auto=pt, vid_stride=vid_stride)
+        bs = len(dataset)
+    elif screenshot:
+        dataset = LoadScreenshots(source, img_size=imgsz, stride=stride, auto=pt)
+    else:
+        dataset = LoadImages(source, img_size=imgsz, stride=stride, auto=pt, vid_stride=vid_stride)
+    vid_path, vid_writer = [None] * bs, [None] * bs
+
+    # Run inference
+    model.warmup(imgsz=(1 if pt else bs, 3, *imgsz))  # warmup
+    seen, windows, dt = 0, [], (Profile(), Profile(), Profile())
+    for path, im, im0s, vid_cap, s in dataset:
+        with dt[0]:
+            im = torch.from_numpy(im).to(model.device)
+            im = im.half() if model.fp16 else im.float()  # uint8 to fp16/32
+            im /= 255  # 0 - 255 to 0.0 - 1.0
+            if len(im.shape) == 3:
+                im = im[None]  # expand for batch dim
+
+        # Inference
+        with dt[1]:
+            visualize = increment_path(save_dir / Path(path).stem, mkdir=True) if visualize else False
+            pred, proto = model(im, augment=augment, visualize=visualize)[:2]
+
+        # NMS
+        with dt[2]:
+            pred = non_max_suppression(pred, conf_thres, iou_thres, classes, agnostic_nms, max_det=max_det, nm=32)
+
+        # Second-stage classifier (optional)
+        # pred = utils.general.apply_classifier(pred, classifier_model, im, im0s)
+
+        # Process predictions
+        for i, det in enumerate(pred):  # per image
+            seen += 1
+            if webcam:  # batch_size >= 1
+                p, im0, frame = path[i], im0s[i].copy(), dataset.count
+                s += f'{i}: '
+            else:
+                p, im0, frame = path, im0s.copy(), getattr(dataset, 'frame', 0)
+
+            p = Path(p)  # to Path
+            save_path = str(save_dir / p.name)  # im.jpg
+            txt_path = str(save_dir / 'labels' / p.stem) + ('' if dataset.mode == 'image' else f'_{frame}')  # im.txt
+            s += '%gx%g ' % im.shape[2:]  # print string
+            imc = im0.copy() if save_crop else im0  # for save_crop
+            annotator = Annotator(im0, line_width=line_thickness, example=str(names))
+            if len(det):
+                masks = process_mask(proto[i], det[:, 6:], det[:, :4], im.shape[2:], upsample=True)  # HWC
+                det[:, :4] = scale_boxes(im.shape[2:], det[:, :4], im0.shape).round()  # rescale boxes to im0 size
+
+                # Segments
+                if save_txt:
+                    segments = reversed(masks2segments(masks))
+                    segments = [scale_segments(im.shape[2:], x, im0.shape, normalize=True) for x in segments]
+
+                # Print results
+                for c in det[:, 5].unique():
+                    n = (det[:, 5] == c).sum()  # detections per class
+                    s += f"{n} {names[int(c)]}{'s' * (n > 1)}, "  # add to string
+
+                # Mask plotting
+                annotator.masks(masks,
+                                colors=[colors(x, True) for x in det[:, 5]],
+                                im_gpu=None if retina_masks else im[i])
+
+                # Write results
+                for j, (*xyxy, conf, cls) in enumerate(reversed(det[:, :6])):
+                    if save_txt:  # Write to file
+                        segj = segments[j].reshape(-1)  # (n,2) to (n*2)
+                        line = (cls, *segj, conf) if save_conf else (cls, *segj)  # label format
+                        with open(f'{txt_path}.txt', 'a') as f:
+                            f.write(('%g ' * len(line)).rstrip() % line + '\n')
+
+                    if save_img or save_crop or view_img:  # Add bbox to image
+                        c = int(cls)  # integer class
+                        label = None if hide_labels else (names[c] if hide_conf else f'{names[c]} {conf:.2f}')
+                        annotator.box_label(xyxy, label, color=colors(c, True))
+                        # annotator.draw.polygon(segments[j], outline=colors(c, True), width=3)
+                    if save_crop:
+                        save_one_box(xyxy, imc, file=save_dir / 'crops' / names[c] / f'{p.stem}.jpg', BGR=True)
+
+            # Stream results
+            im0 = annotator.result()
+            if view_img:
+                if platform.system() == 'Linux' and p not in windows:
+                    windows.append(p)
+                    cv2.namedWindow(str(p), cv2.WINDOW_NORMAL | cv2.WINDOW_KEEPRATIO)  # allow window resize (Linux)
+                    cv2.resizeWindow(str(p), im0.shape[1], im0.shape[0])
+                cv2.imshow(str(p), im0)
+                if cv2.waitKey(1) == ord('q'):  # 1 millisecond
+                    exit()
+
+            # Save results (image with detections)
+            if save_img:
+                if dataset.mode == 'image':
+                    cv2.imwrite(save_path, im0)
+                else:  # 'video' or 'stream'
+                    if vid_path[i] != save_path:  # new video
+                        vid_path[i] = save_path
+                        if isinstance(vid_writer[i], cv2.VideoWriter):
+                            vid_writer[i].release()  # release previous video writer
+                        if vid_cap:  # video
+                            fps = vid_cap.get(cv2.CAP_PROP_FPS)
+                            w = int(vid_cap.get(cv2.CAP_PROP_FRAME_WIDTH))
+                            h = int(vid_cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
+                        else:  # stream
+                            fps, w, h = 30, im0.shape[1], im0.shape[0]
+                        save_path = str(Path(save_path).with_suffix('.mp4'))  # force *.mp4 suffix on results videos
+                        vid_writer[i] = cv2.VideoWriter(save_path, cv2.VideoWriter_fourcc(*'mp4v'), fps, (w, h))
+                    vid_writer[i].write(im0)
+
+        # Print time (inference-only)
+        LOGGER.info(f"{s}{'' if len(det) else '(no detections), '}{dt[1].dt * 1E3:.1f}ms")
+
+    # Print results
+    t = tuple(x.t / seen * 1E3 for x in dt)  # speeds per image
+    LOGGER.info(f'Speed: %.1fms pre-process, %.1fms inference, %.1fms NMS per image at shape {(1, 3, *imgsz)}' % t)
+    if save_txt or save_img:
+        s = f"\n{len(list(save_dir.glob('labels/*.txt')))} labels saved to {save_dir / 'labels'}" if save_txt else ''
+        LOGGER.info(f"Results saved to {colorstr('bold', save_dir)}{s}")
+    if update:
+        strip_optimizer(weights[0])  # update model (to fix SourceChangeWarning)
+
+
+def parse_opt():
+    parser = argparse.ArgumentParser()
+    parser.add_argument('--weights', nargs='+', type=str, default=ROOT / 'yolov5s-seg.pt', help='model path(s)')
+    parser.add_argument('--source', type=str, default=ROOT / 'data/images', help='file/dir/URL/glob/screen/0(webcam)')
+    parser.add_argument('--data', type=str, default=ROOT / 'data/coco128.yaml', help='(optional) dataset.yaml path')
+    parser.add_argument('--imgsz', '--img', '--img-size', nargs='+', type=int, default=[640], help='inference size h,w')
+    parser.add_argument('--conf-thres', type=float, default=0.25, help='confidence threshold')
+    parser.add_argument('--iou-thres', type=float, default=0.45, help='NMS IoU threshold')
+    parser.add_argument('--max-det', type=int, default=1000, help='maximum detections per image')
+    parser.add_argument('--device', default='', help='cuda device, i.e. 0 or 0,1,2,3 or cpu')
+    parser.add_argument('--view-img', action='store_true', help='show results')
+    parser.add_argument('--save-txt', action='store_true', help='save results to *.txt')
+    parser.add_argument('--save-conf', action='store_true', help='save confidences in --save-txt labels')
+    parser.add_argument('--save-crop', action='store_true', help='save cropped prediction boxes')
+    parser.add_argument('--nosave', action='store_true', help='do not save images/videos')
+    parser.add_argument('--classes', nargs='+', type=int, help='filter by class: --classes 0, or --classes 0 2 3')
+    parser.add_argument('--agnostic-nms', action='store_true', help='class-agnostic NMS')
+    parser.add_argument('--augment', action='store_true', help='augmented inference')
+    parser.add_argument('--visualize', action='store_true', help='visualize features')
+    parser.add_argument('--update', action='store_true', help='update all models')
+    parser.add_argument('--project', default=ROOT / 'runs/predict-seg', help='save results to project/name')
+    parser.add_argument('--name', default='exp', help='save results to project/name')
+    parser.add_argument('--exist-ok', action='store_true', help='existing project/name ok, do not increment')
+    parser.add_argument('--line-thickness', default=3, type=int, help='bounding box thickness (pixels)')
+    parser.add_argument('--hide-labels', default=False, action='store_true', help='hide labels')
+    parser.add_argument('--hide-conf', default=False, action='store_true', help='hide confidences')
+    parser.add_argument('--half', action='store_true', help='use FP16 half-precision inference')
+    parser.add_argument('--dnn', action='store_true', help='use OpenCV DNN for ONNX inference')
+    parser.add_argument('--vid-stride', type=int, default=1, help='video frame-rate stride')
+    parser.add_argument('--retina-masks', action='store_true', help='whether to plot masks in native resolution')
+    opt = parser.parse_args()
+    opt.imgsz *= 2 if len(opt.imgsz) == 1 else 1  # expand
+    print_args(vars(opt))
+    return opt
+
+
+def main(opt):
+    check_requirements(exclude=('tensorboard', 'thop'))
+    run(**vars(opt))
+
+
+if __name__ == "__main__":
+    opt = parse_opt()
+    main(opt)

segment/train.py

+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+"""
+Train a YOLOv5 segment model on a segment dataset
+Models and datasets download automatically from the latest YOLOv5 release.
+
+Usage - Single-GPU training:
+    $ python segment/train.py --data coco128-seg.yaml --weights yolov5s-seg.pt --img 640  # from pretrained (recommended)
+    $ python segment/train.py --data coco128-seg.yaml --weights '' --cfg yolov5s-seg.yaml --img 640  # from scratch
+
+Usage - Multi-GPU DDP training:
+    $ python -m torch.distributed.run --nproc_per_node 4 --master_port 1 segment/train.py --data coco128-seg.yaml --weights yolov5s-seg.pt --img 640 --device 0,1,2,3
+
+Models:     https://github.com/ultralytics/yolov5/tree/master/models
+Datasets:   https://github.com/ultralytics/yolov5/tree/master/data
+Tutorial:   https://github.com/ultralytics/yolov5/wiki/Train-Custom-Data
+"""
+
+import argparse
+import math
+import os
+import random
+import sys
+import time
+from copy import deepcopy
+from datetime import datetime
+from pathlib import Path
+
+import numpy as np
+import torch
+import torch.distributed as dist
+import torch.nn as nn
+import yaml
+from torch.optim import lr_scheduler
+from tqdm import tqdm
+
+FILE = Path(__file__).resolve()
+ROOT = FILE.parents[1]  # YOLOv5 root directory
+if str(ROOT) not in sys.path:
+    sys.path.append(str(ROOT))  # add ROOT to PATH
+ROOT = Path(os.path.relpath(ROOT, Path.cwd()))  # relative
+
+import segment.val as validate  # for end-of-epoch mAP
+from models.experimental import attempt_load
+from models.yolo import SegmentationModel
+from utils.autoanchor import check_anchors
+from utils.autobatch import check_train_batch_size
+from utils.callbacks import Callbacks
+from utils.downloads import attempt_download, is_url
+from utils.general import (LOGGER, TQDM_BAR_FORMAT, check_amp, check_dataset, check_file, check_git_info,
+                           check_git_status, check_img_size, check_requirements, check_suffix, check_yaml, colorstr,
+                           get_latest_run, increment_path, init_seeds, intersect_dicts, labels_to_class_weights,
+                           labels_to_image_weights, one_cycle, print_args, print_mutation, strip_optimizer, yaml_save)
+from utils.loggers import GenericLogger
+from utils.plots import plot_evolve, plot_labels
+from utils.segment.dataloaders import create_dataloader
+from utils.segment.loss import ComputeLoss
+from utils.segment.metrics import KEYS, fitness
+from utils.segment.plots import plot_images_and_masks, plot_results_with_masks
+from utils.torch_utils import (EarlyStopping, ModelEMA, de_parallel, select_device, smart_DDP, smart_optimizer,
+                               smart_resume, torch_distributed_zero_first)
+
+LOCAL_RANK = int(os.getenv('LOCAL_RANK', -1))  # https://pytorch.org/docs/stable/elastic/run.html
+RANK = int(os.getenv('RANK', -1))
+WORLD_SIZE = int(os.getenv('WORLD_SIZE', 1))
+GIT_INFO = check_git_info()
+
+
+def train(hyp, opt, device, callbacks):  # hyp is path/to/hyp.yaml or hyp dictionary
+    save_dir, epochs, batch_size, weights, single_cls, evolve, data, cfg, resume, noval, nosave, workers, freeze, mask_ratio = \
+        Path(opt.save_dir), opt.epochs, opt.batch_size, opt.weights, opt.single_cls, opt.evolve, opt.data, opt.cfg, \
+        opt.resume, opt.noval, opt.nosave, opt.workers, opt.freeze, opt.mask_ratio
+    # callbacks.run('on_pretrain_routine_start')
+
+    # Directories
+    w = save_dir / 'weights'  # weights dir
+    (w.parent if evolve else w).mkdir(parents=True, exist_ok=True)  # make dir
+    last, best = w / 'last.pt', w / 'best.pt'
+
+    # Hyperparameters
+    if isinstance(hyp, str):
+        with open(hyp, errors='ignore') as f:
+            hyp = yaml.safe_load(f)  # load hyps dict
+    LOGGER.info(colorstr('hyperparameters: ') + ', '.join(f'{k}={v}' for k, v in hyp.items()))
+    opt.hyp = hyp.copy()  # for saving hyps to checkpoints
+
+    # Save run settings
+    if not evolve:
+        yaml_save(save_dir / 'hyp.yaml', hyp)
+        yaml_save(save_dir / 'opt.yaml', vars(opt))
+
+    # Loggers
+    data_dict = None
+    if RANK in {-1, 0}:
+        logger = GenericLogger(opt=opt, console_logger=LOGGER)
+
+    # Config
+    plots = not evolve and not opt.noplots  # create plots
+    overlap = not opt.no_overlap
+    cuda = device.type != 'cpu'
+    init_seeds(opt.seed + 1 + RANK, deterministic=True)
+    with torch_distributed_zero_first(LOCAL_RANK):
+        data_dict = data_dict or check_dataset(data)  # check if None
+    train_path, val_path = data_dict['train'], data_dict['val']
+    nc = 1 if single_cls else int(data_dict['nc'])  # number of classes
+    names = {0: 'item'} if single_cls and len(data_dict['names']) != 1 else data_dict['names']  # class names
+    is_coco = isinstance(val_path, str) and val_path.endswith('coco/val2017.txt')  # COCO dataset
+
+    # Model
+    check_suffix(weights, '.pt')  # check weights
+    pretrained = weights.endswith('.pt')
+    if pretrained:
+        with torch_distributed_zero_first(LOCAL_RANK):
+            weights = attempt_download(weights)  # download if not found locally
+        ckpt = torch.load(weights, map_location='cpu')  # load checkpoint to CPU to avoid CUDA memory leak
+        model = SegmentationModel(cfg or ckpt['model'].yaml, ch=3, nc=nc, anchors=hyp.get('anchors')).to(device)
+        exclude = ['anchor'] if (cfg or hyp.get('anchors')) and not resume else []  # exclude keys
+        csd = ckpt['model'].float().state_dict()  # checkpoint state_dict as FP32
+        csd = intersect_dicts(csd, model.state_dict(), exclude=exclude)  # intersect
+        model.load_state_dict(csd, strict=False)  # load
+        LOGGER.info(f'Transferred {len(csd)}/{len(model.state_dict())} items from {weights}')  # report
+    else:
+        model = SegmentationModel(cfg, ch=3, nc=nc, anchors=hyp.get('anchors')).to(device)  # create
+    amp = check_amp(model)  # check AMP
+
+    # Freeze
+    freeze = [f'model.{x}.' for x in (freeze if len(freeze) > 1 else range(freeze[0]))]  # layers to freeze
+    for k, v in model.named_parameters():
+        v.requires_grad = True  # train all layers
+        # v.register_hook(lambda x: torch.nan_to_num(x))  # NaN to 0 (commented for erratic training results)
+        if any(x in k for x in freeze):
+            LOGGER.info(f'freezing {k}')
+            v.requires_grad = False
+
+    # Image size
+    gs = max(int(model.stride.max()), 32)  # grid size (max stride)
+    imgsz = check_img_size(opt.imgsz, gs, floor=gs * 2)  # verify imgsz is gs-multiple
+
+    # Batch size
+    if RANK == -1 and batch_size == -1:  # single-GPU only, estimate best batch size
+        batch_size = check_train_batch_size(model, imgsz, amp)
+        logger.update_params({"batch_size": batch_size})
+        # loggers.on_params_update({"batch_size": batch_size})
+
+    # Optimizer
+    nbs = 64  # nominal batch size
+    accumulate = max(round(nbs / batch_size), 1)  # accumulate loss before optimizing
+    hyp['weight_decay'] *= batch_size * accumulate / nbs  # scale weight_decay
+    optimizer = smart_optimizer(model, opt.optimizer, hyp['lr0'], hyp['momentum'], hyp['weight_decay'])
+
+    # Scheduler
+    if opt.cos_lr:
+        lf = one_cycle(1, hyp['lrf'], epochs)  # cosine 1->hyp['lrf']
+    else:
+        lf = lambda x: (1 - x / epochs) * (1.0 - hyp['lrf']) + hyp['lrf']  # linear
+    scheduler = lr_scheduler.LambdaLR(optimizer, lr_lambda=lf)  # plot_lr_scheduler(optimizer, scheduler, epochs)
+
+    # EMA
+    ema = ModelEMA(model) if RANK in {-1, 0} else None
+
+    # Resume
+    best_fitness, start_epoch = 0.0, 0
+    if pretrained:
+        if resume:
+            best_fitness, start_epoch, epochs = smart_resume(ckpt, optimizer, ema, weights, epochs, resume)
+        del ckpt, csd
+
+    # DP mode
+    if cuda and RANK == -1 and torch.cuda.device_count() > 1:
+        LOGGER.warning('WARNING ⚠️ DP not recommended, use torch.distributed.run for best DDP Multi-GPU results.\n'
+                       'See Multi-GPU Tutorial at https://github.com/ultralytics/yolov5/issues/475 to get started.')
+        model = torch.nn.DataParallel(model)
+
+    # SyncBatchNorm
+    if opt.sync_bn and cuda and RANK != -1:
+        model = torch.nn.SyncBatchNorm.convert_sync_batchnorm(model).to(device)
+        LOGGER.info('Using SyncBatchNorm()')
+
+    # Trainloader
+    train_loader, dataset = create_dataloader(
+        train_path,
+        imgsz,
+        batch_size // WORLD_SIZE,
+        gs,
+        single_cls,
+        hyp=hyp,
+        augment=True,
+        cache=None if opt.cache == 'val' else opt.cache,
+        rect=opt.rect,
+        rank=LOCAL_RANK,
+        workers=workers,
+        image_weights=opt.image_weights,
+        quad=opt.quad,
+        prefix=colorstr('train: '),
+        shuffle=True,
+        mask_downsample_ratio=mask_ratio,
+        overlap_mask=overlap,
+    )
+    labels = np.concatenate(dataset.labels, 0)
+    mlc = int(labels[:, 0].max())  # max label class
+    assert mlc < nc, f'Label class {mlc} exceeds nc={nc} in {data}. Possible class labels are 0-{nc - 1}'
+
+    # Process 0
+    if RANK in {-1, 0}:
+        val_loader = create_dataloader(val_path,
+                                       imgsz,
+                                       batch_size // WORLD_SIZE * 2,
+                                       gs,
+                                       single_cls,
+                                       hyp=hyp,
+                                       cache=None if noval else opt.cache,
+                                       rect=True,
+                                       rank=-1,
+                                       workers=workers * 2,
+                                       pad=0.5,
+                                       mask_downsample_ratio=mask_ratio,
+                                       overlap_mask=overlap,
+                                       prefix=colorstr('val: '))[0]
+
+        if not resume:
+            if not opt.noautoanchor:
+                check_anchors(dataset, model=model, thr=hyp['anchor_t'], imgsz=imgsz)  # run AutoAnchor
+            model.half().float()  # pre-reduce anchor precision
+
+            if plots:
+                plot_labels(labels, names, save_dir)
+        # callbacks.run('on_pretrain_routine_end', labels, names)
+
+    # DDP mode
+    if cuda and RANK != -1:
+        model = smart_DDP(model)
+
+    # Model attributes
+    nl = de_parallel(model).model[-1].nl  # number of detection layers (to scale hyps)
+    hyp['box'] *= 3 / nl  # scale to layers
+    hyp['cls'] *= nc / 80 * 3 / nl  # scale to classes and layers
+    hyp['obj'] *= (imgsz / 640) ** 2 * 3 / nl  # scale to image size and layers
+    hyp['label_smoothing'] = opt.label_smoothing
+    model.nc = nc  # attach number of classes to model
+    model.hyp = hyp  # attach hyperparameters to model
+    model.class_weights = labels_to_class_weights(dataset.labels, nc).to(device) * nc  # attach class weights
+    model.names = names
+
+    # Start training
+    t0 = time.time()
+    nb = len(train_loader)  # number of batches
+    nw = max(round(hyp['warmup_epochs'] * nb), 100)  # number of warmup iterations, max(3 epochs, 100 iterations)
+    # nw = min(nw, (epochs - start_epoch) / 2 * nb)  # limit warmup to < 1/2 of training
+    last_opt_step = -1
+    maps = np.zeros(nc)  # mAP per class
+    results = (0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0)  # P, R, mAP@.5, mAP@.5-.95, val_loss(box, obj, cls)
+    scheduler.last_epoch = start_epoch - 1  # do not move
+    scaler = torch.cuda.amp.GradScaler(enabled=amp)
+    stopper, stop = EarlyStopping(patience=opt.patience), False
+    compute_loss = ComputeLoss(model, overlap=overlap)  # init loss class
+    # callbacks.run('on_train_start')
+    LOGGER.info(f'Image sizes {imgsz} train, {imgsz} val\n'
+                f'Using {train_loader.num_workers * WORLD_SIZE} dataloader workers\n'
+                f"Logging results to {colorstr('bold', save_dir)}\n"
+                f'Starting training for {epochs} epochs...')
+    for epoch in range(start_epoch, epochs):  # epoch ------------------------------------------------------------------
+        # callbacks.run('on_train_epoch_start')
+        model.train()
+
+        # Update image weights (optional, single-GPU only)
+        if opt.image_weights:
+            cw = model.class_weights.cpu().numpy() * (1 - maps) ** 2 / nc  # class weights
+            iw = labels_to_image_weights(dataset.labels, nc=nc, class_weights=cw)  # image weights
+            dataset.indices = random.choices(range(dataset.n), weights=iw, k=dataset.n)  # rand weighted idx
+
+        # Update mosaic border (optional)
+        # b = int(random.uniform(0.25 * imgsz, 0.75 * imgsz + gs) // gs * gs)
+        # dataset.mosaic_border = [b - imgsz, -b]  # height, width borders
+
+        mloss = torch.zeros(4, device=device)  # mean losses
+        if RANK != -1:
+            train_loader.sampler.set_epoch(epoch)
+        pbar = enumerate(train_loader)
+        LOGGER.info(('\n' + '%11s' * 8) %
+                    ('Epoch', 'GPU_mem', 'box_loss', 'seg_loss', 'obj_loss', 'cls_loss', 'Instances', 'Size'))
+        if RANK in {-1, 0}:
+            pbar = tqdm(pbar, total=nb, bar_format=TQDM_BAR_FORMAT)  # progress bar
+        optimizer.zero_grad()
+        for i, (imgs, targets, paths, _, masks) in pbar:  # batch ------------------------------------------------------
+            # callbacks.run('on_train_batch_start')
+            ni = i + nb * epoch  # number integrated batches (since train start)
+            imgs = imgs.to(device, non_blocking=True).float() / 255  # uint8 to float32, 0-255 to 0.0-1.0
+
+            # Warmup
+            if ni <= nw:
+                xi = [0, nw]  # x interp
+                # compute_loss.gr = np.interp(ni, xi, [0.0, 1.0])  # iou loss ratio (obj_loss = 1.0 or iou)
+                accumulate = max(1, np.interp(ni, xi, [1, nbs / batch_size]).round())
+                for j, x in enumerate(optimizer.param_groups):
+                    # bias lr falls from 0.1 to lr0, all other lrs rise from 0.0 to lr0
+                    x['lr'] = np.interp(ni, xi, [hyp['warmup_bias_lr'] if j == 0 else 0.0, x['initial_lr'] * lf(epoch)])
+                    if 'momentum' in x:
+                        x['momentum'] = np.interp(ni, xi, [hyp['warmup_momentum'], hyp['momentum']])
+
+            # Multi-scale
+            if opt.multi_scale:
+                sz = random.randrange(imgsz * 0.5, imgsz * 1.5 + gs) // gs * gs  # size
+                sf = sz / max(imgs.shape[2:])  # scale factor
+                if sf != 1:
+                    ns = [math.ceil(x * sf / gs) * gs for x in imgs.shape[2:]]  # new shape (stretched to gs-multiple)
+                    imgs = nn.functional.interpolate(imgs, size=ns, mode='bilinear', align_corners=False)
+
+            # Forward
+            with torch.cuda.amp.autocast(amp):
+                pred = model(imgs)  # forward
+                loss, loss_items = compute_loss(pred, targets.to(device), masks=masks.to(device).float())
+                if RANK != -1:
+                    loss *= WORLD_SIZE  # gradient averaged between devices in DDP mode
+                if opt.quad:
+                    loss *= 4.
+
+            # Backward
+            scaler.scale(loss).backward()
+
+            # Optimize - https://pytorch.org/docs/master/notes/amp_examples.html
+            if ni - last_opt_step >= accumulate:
+                scaler.unscale_(optimizer)  # unscale gradients
+                torch.nn.utils.clip_grad_norm_(model.parameters(), max_norm=10.0)  # clip gradients
+                scaler.step(optimizer)  # optimizer.step
+                scaler.update()
+                optimizer.zero_grad()
+                if ema:
+                    ema.update(model)
+                last_opt_step = ni
+
+            # Log
+            if RANK in {-1, 0}:
+                mloss = (mloss * i + loss_items) / (i + 1)  # update mean losses
+                mem = f'{torch.cuda.memory_reserved() / 1E9 if torch.cuda.is_available() else 0:.3g}G'  # (GB)
+                pbar.set_description(('%11s' * 2 + '%11.4g' * 6) %
+                                     (f'{epoch}/{epochs - 1}', mem, *mloss, targets.shape[0], imgs.shape[-1]))
+                # callbacks.run('on_train_batch_end', model, ni, imgs, targets, paths)
+                # if callbacks.stop_training:
+                #    return
+
+                # Mosaic plots
+                if plots:
+                    if ni < 3:
+                        plot_images_and_masks(imgs, targets, masks, paths, save_dir / f"train_batch{ni}.jpg")
+                    if ni == 10:
+                        files = sorted(save_dir.glob('train*.jpg'))
+                        logger.log_images(files, "Mosaics", epoch)
+            # end batch ------------------------------------------------------------------------------------------------
+
+        # Scheduler
+        lr = [x['lr'] for x in optimizer.param_groups]  # for loggers
+        scheduler.step()
+
+        if RANK in {-1, 0}:
+            # mAP
+            # callbacks.run('on_train_epoch_end', epoch=epoch)
+            ema.update_attr(model, include=['yaml', 'nc', 'hyp', 'names', 'stride', 'class_weights'])
+            final_epoch = (epoch + 1 == epochs) or stopper.possible_stop
+            if not noval or final_epoch:  # Calculate mAP
+                results, maps, _ = validate.run(data_dict,
+                                                batch_size=batch_size // WORLD_SIZE * 2,
+                                                imgsz=imgsz,
+                                                half=amp,
+                                                model=ema.ema,
+                                                single_cls=single_cls,
+                                                dataloader=val_loader,
+                                                save_dir=save_dir,
+                                                plots=False,
+                                                callbacks=callbacks,
+                                                compute_loss=compute_loss,
+                                                mask_downsample_ratio=mask_ratio,
+                                                overlap=overlap)
+
+            # Update best mAP
+            fi = fitness(np.array(results).reshape(1, -1))  # weighted combination of [P, R, mAP@.5, mAP@.5-.95]
+            stop = stopper(epoch=epoch, fitness=fi)  # early stop check
+            if fi > best_fitness:
+                best_fitness = fi
+            log_vals = list(mloss) + list(results) + lr
+            # callbacks.run('on_fit_epoch_end', log_vals, epoch, best_fitness, fi)
+            # Log val metrics and media
+            metrics_dict = dict(zip(KEYS, log_vals))
+            logger.log_metrics(metrics_dict, epoch)
+
+            # Save model
+            if (not nosave) or (final_epoch and not evolve):  # if save
+                ckpt = {
+                    'epoch': epoch,
+                    'best_fitness': best_fitness,
+                    'model': deepcopy(de_parallel(model)).half(),
+                    'ema': deepcopy(ema.ema).half(),
+                    'updates': ema.updates,
+                    'optimizer': optimizer.state_dict(),
+                    'opt': vars(opt),
+                    'git': GIT_INFO,  # {remote, branch, commit} if a git repo
+                    'date': datetime.now().isoformat()}
+
+                # Save last, best and delete
+                torch.save(ckpt, last)
+                if best_fitness == fi:
+                    torch.save(ckpt, best)
+                if opt.save_period > 0 and epoch % opt.save_period == 0:
+                    torch.save(ckpt, w / f'epoch{epoch}.pt')
+                    logger.log_model(w / f'epoch{epoch}.pt')
+                del ckpt
+                # callbacks.run('on_model_save', last, epoch, final_epoch, best_fitness, fi)
+
+        # EarlyStopping
+        if RANK != -1:  # if DDP training
+            broadcast_list = [stop if RANK == 0 else None]
+            dist.broadcast_object_list(broadcast_list, 0)  # broadcast 'stop' to all ranks
+            if RANK != 0:
+                stop = broadcast_list[0]
+        if stop:
+            break  # must break all DDP ranks
+
+        # end epoch ----------------------------------------------------------------------------------------------------
+    # end training -----------------------------------------------------------------------------------------------------
+    if RANK in {-1, 0}:
+        LOGGER.info(f'\n{epoch - start_epoch + 1} epochs completed in {(time.time() - t0) / 3600:.3f} hours.')
+        for f in last, best:
+            if f.exists():
+                strip_optimizer(f)  # strip optimizers
+                if f is best:
+                    LOGGER.info(f'\nValidating {f}...')
+                    results, _, _ = validate.run(
+                        data_dict,
+                        batch_size=batch_size // WORLD_SIZE * 2,
+                        imgsz=imgsz,
+                        model=attempt_load(f, device).half(),
+                        iou_thres=0.65 if is_coco else 0.60,  # best pycocotools at iou 0.65
+                        single_cls=single_cls,
+                        dataloader=val_loader,
+                        save_dir=save_dir,
+                        save_json=is_coco,
+                        verbose=True,
+                        plots=plots,
+                        callbacks=callbacks,
+                        compute_loss=compute_loss,
+                        mask_downsample_ratio=mask_ratio,
+                        overlap=overlap)  # val best model with plots
+                    if is_coco:
+                        # callbacks.run('on_fit_epoch_end', list(mloss) + list(results) + lr, epoch, best_fitness, fi)
+                        metrics_dict = dict(zip(KEYS, list(mloss) + list(results) + lr))
+                        logger.log_metrics(metrics_dict, epoch)
+
+        # callbacks.run('on_train_end', last, best, epoch, results)
+        # on train end callback using genericLogger
+        logger.log_metrics(dict(zip(KEYS[4:16], results)), epochs)
+        if not opt.evolve:
+            logger.log_model(best, epoch)
+        if plots:
+            plot_results_with_masks(file=save_dir / 'results.csv')  # save results.png
+            files = ['results.png', 'confusion_matrix.png', *(f'{x}_curve.png' for x in ('F1', 'PR', 'P', 'R'))]
+            files = [(save_dir / f) for f in files if (save_dir / f).exists()]  # filter
+            LOGGER.info(f"Results saved to {colorstr('bold', save_dir)}")
+            logger.log_images(files, "Results", epoch + 1)
+            logger.log_images(sorted(save_dir.glob('val*.jpg')), "Validation", epoch + 1)
+    torch.cuda.empty_cache()
+    return results
+
+
+def parse_opt(known=False):
+    parser = argparse.ArgumentParser()
+    parser.add_argument('--weights', type=str, default=ROOT / 'yolov5s-seg.pt', help='initial weights path')
+    parser.add_argument('--cfg', type=str, default='', help='model.yaml path')
+    parser.add_argument('--data', type=str, default=ROOT / 'data/coco128-seg.yaml', help='dataset.yaml path')
+    parser.add_argument('--hyp', type=str, default=ROOT / 'data/hyps/hyp.scratch-low.yaml', help='hyperparameters path')
+    parser.add_argument('--epochs', type=int, default=100, help='total training epochs')
+    parser.add_argument('--batch-size', type=int, default=16, help='total batch size for all GPUs, -1 for autobatch')
+    parser.add_argument('--imgsz', '--img', '--img-size', type=int, default=640, help='train, val image size (pixels)')
+    parser.add_argument('--rect', action='store_true', help='rectangular training')
+    parser.add_argument('--resume', nargs='?', const=True, default=False, help='resume most recent training')
+    parser.add_argument('--nosave', action='store_true', help='only save final checkpoint')
+    parser.add_argument('--noval', action='store_true', help='only validate final epoch')
+    parser.add_argument('--noautoanchor', action='store_true', help='disable AutoAnchor')
+    parser.add_argument('--noplots', action='store_true', help='save no plot files')
+    parser.add_argument('--evolve', type=int, nargs='?', const=300, help='evolve hyperparameters for x generations')
+    parser.add_argument('--bucket', type=str, default='', help='gsutil bucket')
+    parser.add_argument('--cache', type=str, nargs='?', const='ram', help='image --cache ram/disk')
+    parser.add_argument('--image-weights', action='store_true', help='use weighted image selection for training')
+    parser.add_argument('--device', default='', help='cuda device, i.e. 0 or 0,1,2,3 or cpu')
+    parser.add_argument('--multi-scale', action='store_true', help='vary img-size +/- 50%%')
+    parser.add_argument('--single-cls', action='store_true', help='train multi-class data as single-class')
+    parser.add_argument('--optimizer', type=str, choices=['SGD', 'Adam', 'AdamW'], default='SGD', help='optimizer')
+    parser.add_argument('--sync-bn', action='store_true', help='use SyncBatchNorm, only available in DDP mode')
+    parser.add_argument('--workers', type=int, default=8, help='max dataloader workers (per RANK in DDP mode)')
+    parser.add_argument('--project', default=ROOT / 'runs/train-seg', help='save to project/name')
+    parser.add_argument('--name', default='exp', help='save to project/name')
+    parser.add_argument('--exist-ok', action='store_true', help='existing project/name ok, do not increment')
+    parser.add_argument('--quad', action='store_true', help='quad dataloader')
+    parser.add_argument('--cos-lr', action='store_true', help='cosine LR scheduler')
+    parser.add_argument('--label-smoothing', type=float, default=0.0, help='Label smoothing epsilon')
+    parser.add_argument('--patience', type=int, default=100, help='EarlyStopping patience (epochs without improvement)')
+    parser.add_argument('--freeze', nargs='+', type=int, default=[0], help='Freeze layers: backbone=10, first3=0 1 2')
+    parser.add_argument('--save-period', type=int, default=-1, help='Save checkpoint every x epochs (disabled if < 1)')
+    parser.add_argument('--seed', type=int, default=0, help='Global training seed')
+    parser.add_argument('--local_rank', type=int, default=-1, help='Automatic DDP Multi-GPU argument, do not modify')
+
+    # Instance Segmentation Args
+    parser.add_argument('--mask-ratio', type=int, default=4, help='Downsample the truth masks to saving memory')
+    parser.add_argument('--no-overlap', action='store_true', help='Overlap masks train faster at slightly less mAP')
+
+    return parser.parse_known_args()[0] if known else parser.parse_args()
+
+
+def main(opt, callbacks=Callbacks()):
+    # Checks
+    if RANK in {-1, 0}:
+        print_args(vars(opt))
+        check_git_status()
+        check_requirements()
+
+    # Resume
+    if opt.resume and not opt.evolve:  # resume from specified or most recent last.pt
+        last = Path(check_file(opt.resume) if isinstance(opt.resume, str) else get_latest_run())
+        opt_yaml = last.parent.parent / 'opt.yaml'  # train options yaml
+        opt_data = opt.data  # original dataset
+        if opt_yaml.is_file():
+            with open(opt_yaml, errors='ignore') as f:
+                d = yaml.safe_load(f)
+        else:
+            d = torch.load(last, map_location='cpu')['opt']
+        opt = argparse.Namespace(**d)  # replace
+        opt.cfg, opt.weights, opt.resume = '', str(last), True  # reinstate
+        if is_url(opt_data):
+            opt.data = check_file(opt_data)  # avoid HUB resume auth timeout
+    else:
+        opt.data, opt.cfg, opt.hyp, opt.weights, opt.project = \
+            check_file(opt.data), check_yaml(opt.cfg), check_yaml(opt.hyp), str(opt.weights), str(opt.project)  # checks
+        assert len(opt.cfg) or len(opt.weights), 'either --cfg or --weights must be specified'
+        if opt.evolve:
+            if opt.project == str(ROOT / 'runs/train'):  # if default project name, rename to runs/evolve
+                opt.project = str(ROOT / 'runs/evolve')
+            opt.exist_ok, opt.resume = opt.resume, False  # pass resume to exist_ok and disable resume
+        if opt.name == 'cfg':
+            opt.name = Path(opt.cfg).stem  # use model.yaml as name
+        opt.save_dir = str(increment_path(Path(opt.project) / opt.name, exist_ok=opt.exist_ok))
+
+    # DDP mode
+    device = select_device(opt.device, batch_size=opt.batch_size)
+    if LOCAL_RANK != -1:
+        msg = 'is not compatible with YOLOv5 Multi-GPU DDP training'
+        assert not opt.image_weights, f'--image-weights {msg}'
+        assert not opt.evolve, f'--evolve {msg}'
+        assert opt.batch_size != -1, f'AutoBatch with --batch-size -1 {msg}, please pass a valid --batch-size'
+        assert opt.batch_size % WORLD_SIZE == 0, f'--batch-size {opt.batch_size} must be multiple of WORLD_SIZE'
+        assert torch.cuda.device_count() > LOCAL_RANK, 'insufficient CUDA devices for DDP command'
+        torch.cuda.set_device(LOCAL_RANK)
+        device = torch.device('cuda', LOCAL_RANK)
+        dist.init_process_group(backend="nccl" if dist.is_nccl_available() else "gloo")
+
+    # Train
+    if not opt.evolve:
+        train(opt.hyp, opt, device, callbacks)
+
+    # Evolve hyperparameters (optional)
+    else:
+        # Hyperparameter evolution metadata (mutation scale 0-1, lower_limit, upper_limit)
+        meta = {
+            'lr0': (1, 1e-5, 1e-1),  # initial learning rate (SGD=1E-2, Adam=1E-3)
+            'lrf': (1, 0.01, 1.0),  # final OneCycleLR learning rate (lr0 * lrf)
+            'momentum': (0.3, 0.6, 0.98),  # SGD momentum/Adam beta1
+            'weight_decay': (1, 0.0, 0.001),  # optimizer weight decay
+            'warmup_epochs': (1, 0.0, 5.0),  # warmup epochs (fractions ok)
+            'warmup_momentum': (1, 0.0, 0.95),  # warmup initial momentum
+            'warmup_bias_lr': (1, 0.0, 0.2),  # warmup initial bias lr
+            'box': (1, 0.02, 0.2),  # box loss gain
+            'cls': (1, 0.2, 4.0),  # cls loss gain
+            'cls_pw': (1, 0.5, 2.0),  # cls BCELoss positive_weight
+            'obj': (1, 0.2, 4.0),  # obj loss gain (scale with pixels)
+            'obj_pw': (1, 0.5, 2.0),  # obj BCELoss positive_weight
+            'iou_t': (0, 0.1, 0.7),  # IoU training threshold
+            'anchor_t': (1, 2.0, 8.0),  # anchor-multiple threshold
+            'anchors': (2, 2.0, 10.0),  # anchors per output grid (0 to ignore)
+            'fl_gamma': (0, 0.0, 2.0),  # focal loss gamma (efficientDet default gamma=1.5)
+            'hsv_h': (1, 0.0, 0.1),  # image HSV-Hue augmentation (fraction)
+            'hsv_s': (1, 0.0, 0.9),  # image HSV-Saturation augmentation (fraction)
+            'hsv_v': (1, 0.0, 0.9),  # image HSV-Value augmentation (fraction)
+            'degrees': (1, 0.0, 45.0),  # image rotation (+/- deg)
+            'translate': (1, 0.0, 0.9),  # image translation (+/- fraction)
+            'scale': (1, 0.0, 0.9),  # image scale (+/- gain)
+            'shear': (1, 0.0, 10.0),  # image shear (+/- deg)
+            'perspective': (0, 0.0, 0.001),  # image perspective (+/- fraction), range 0-0.001
+            'flipud': (1, 0.0, 1.0),  # image flip up-down (probability)
+            'fliplr': (0, 0.0, 1.0),  # image flip left-right (probability)
+            'mosaic': (1, 0.0, 1.0),  # image mixup (probability)
+            'mixup': (1, 0.0, 1.0),  # image mixup (probability)
+            'copy_paste': (1, 0.0, 1.0)}  # segment copy-paste (probability)
+
+        with open(opt.hyp, errors='ignore') as f:
+            hyp = yaml.safe_load(f)  # load hyps dict
+            if 'anchors' not in hyp:  # anchors commented in hyp.yaml
+                hyp['anchors'] = 3
+        if opt.noautoanchor:
+            del hyp['anchors'], meta['anchors']
+        opt.noval, opt.nosave, save_dir = True, True, Path(opt.save_dir)  # only val/save final epoch
+        # ei = [isinstance(x, (int, float)) for x in hyp.values()]  # evolvable indices
+        evolve_yaml, evolve_csv = save_dir / 'hyp_evolve.yaml', save_dir / 'evolve.csv'
+        if opt.bucket:
+            os.system(f'gsutil cp gs://{opt.bucket}/evolve.csv {evolve_csv}')  # download evolve.csv if exists
+
+        for _ in range(opt.evolve):  # generations to evolve
+            if evolve_csv.exists():  # if evolve.csv exists: select best hyps and mutate
+                # Select parent(s)
+                parent = 'single'  # parent selection method: 'single' or 'weighted'
+                x = np.loadtxt(evolve_csv, ndmin=2, delimiter=',', skiprows=1)
+                n = min(5, len(x))  # number of previous results to consider
+                x = x[np.argsort(-fitness(x))][:n]  # top n mutations
+                w = fitness(x) - fitness(x).min() + 1E-6  # weights (sum > 0)
+                if parent == 'single' or len(x) == 1:
+                    # x = x[random.randint(0, n - 1)]  # random selection
+                    x = x[random.choices(range(n), weights=w)[0]]  # weighted selection
+                elif parent == 'weighted':
+                    x = (x * w.reshape(n, 1)).sum(0) / w.sum()  # weighted combination
+
+                # Mutate
+                mp, s = 0.8, 0.2  # mutation probability, sigma
+                npr = np.random
+                npr.seed(int(time.time()))
+                g = np.array([meta[k][0] for k in hyp.keys()])  # gains 0-1
+                ng = len(meta)
+                v = np.ones(ng)
+                while all(v == 1):  # mutate until a change occurs (prevent duplicates)
+                    v = (g * (npr.random(ng) < mp) * npr.randn(ng) * npr.random() * s + 1).clip(0.3, 3.0)
+                for i, k in enumerate(hyp.keys()):  # plt.hist(v.ravel(), 300)
+                    hyp[k] = float(x[i + 7] * v[i])  # mutate
+
+            # Constrain to limits
+            for k, v in meta.items():
+                hyp[k] = max(hyp[k], v[1])  # lower limit
+                hyp[k] = min(hyp[k], v[2])  # upper limit
+                hyp[k] = round(hyp[k], 5)  # significant digits
+
+            # Train mutation
+            results = train(hyp.copy(), opt, device, callbacks)
+            callbacks = Callbacks()
+            # Write mutation results
+            print_mutation(KEYS, results, hyp.copy(), save_dir, opt.bucket)
+
+        # Plot results
+        plot_evolve(evolve_csv)
+        LOGGER.info(f'Hyperparameter evolution finished {opt.evolve} generations\n'
+                    f"Results saved to {colorstr('bold', save_dir)}\n"
+                    f'Usage example: $ python train.py --hyp {evolve_yaml}')
+
+
+def run(**kwargs):
+    # Usage: import train; train.run(data='coco128.yaml', imgsz=320, weights='yolov5m.pt')
+    opt = parse_opt(True)
+    for k, v in kwargs.items():
+        setattr(opt, k, v)
+    main(opt)
+    return opt
+
+
+if __name__ == "__main__":
+    opt = parse_opt()
+    main(opt)

segment/tutorial.ipynb

+{
+  "cells": [
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "t6MPjfT5NrKQ"
+      },
+      "source": [
+        "<div align=\"center\">\n",
+        "\n",
+        "  <a href=\"https://ultralytics.com/yolov5\" target=\"_blank\">\n",
+        "    <img width=\"1024\", src=\"https://raw.githubusercontent.com/ultralytics/assets/master/yolov5/v70/splash.png\"></a>\n",
+        "\n",
+        "\n",
+        "<br>\n",
+        "  <a href=\"https://bit.ly/yolov5-paperspace-notebook\"><img src=\"https://assets.paperspace.io/img/gradient-badge.svg\" alt=\"Run on Gradient\"></a>\n",
+        "  <a href=\"https://colab.research.google.com/github/ultralytics/yolov5/blob/master/segment/tutorial.ipynb\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"></a>\n",
+        "  <a href=\"https://www.kaggle.com/ultralytics/yolov5\"><img src=\"https://kaggle.com/static/images/open-in-kaggle.svg\" alt=\"Open In Kaggle\"></a>\n",
+        "<br>\n",
+        "\n",
+        "This <a href=\"https://github.com/ultralytics/yolov5\">YOLOv5</a> 🚀 notebook by <a href=\"https://ultralytics.com\">Ultralytics</a> presents simple train, validate and predict examples to help start your AI adventure.<br>See <a href=\"https://github.com/ultralytics/yolov5/issues/new/choose\">GitHub</a> for community support or <a href=\"https://ultralytics.com/contact\">contact us</a> for professional support.\n",
+        "\n",
+        "</div>"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "7mGmQbAO5pQb"
+      },
+      "source": [
+        "# Setup\n",
+        "\n",
+        "Clone GitHub [repository](https://github.com/ultralytics/yolov5), install [dependencies](https://github.com/ultralytics/yolov5/blob/master/requirements.txt) and check PyTorch and GPU."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 1,
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "wbvMlHd_QwMG",
+        "outputId": "664f49fa-554a-4dca-8d0e-5c9dd60f6d28"
+      },
+      "outputs": [
+        {
+          "output_type": "stream",
+          "name": "stderr",
+          "text": [
+            "YOLOv5 🚀 v6.2-257-g2ecaa96 Python-3.7.15 torch-1.12.1+cu113 CUDA:0 (Tesla T4, 15110MiB)\n"
+          ]
+        },
+        {
+          "output_type": "stream",
+          "name": "stdout",
+          "text": [
+            "Setup complete ✅ (2 CPUs, 12.7 GB RAM, 22.6/78.2 GB disk)\n"
+          ]
+        }
+      ],
+      "source": [
+        "!git clone https://github.com/ultralytics/yolov5  # clone\n",
+        "%cd yolov5\n",
+        "%pip install -qr requirements.txt  # install\n",
+        "\n",
+        "import torch\n",
+        "import utils\n",
+        "display = utils.notebook_init()  # checks"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "4JnkELT0cIJg"
+      },
+      "source": [
+        "# 1. Predict\n",
+        "\n",
+        "`segment/predict.py` runs YOLOv5 instance segmentation inference on a variety of sources, downloading models automatically from the [latest YOLOv5 release](https://github.com/ultralytics/yolov5/releases), and saving results to `runs/predict`. Example inference sources are:\n",
+        "\n",
+        "```shell\n",
+        "python segment/predict.py --source 0  # webcam\n",
+        "                             img.jpg  # image \n",
+        "                             vid.mp4  # video\n",
+        "                             screen  # screenshot\n",
+        "                             path/  # directory\n",
+        "                             'path/*.jpg'  # glob\n",
+        "                             'https://youtu.be/Zgi9g1ksQHc'  # YouTube\n",
+        "                             'rtsp://example.com/media.mp4'  # RTSP, RTMP, HTTP stream\n",
+        "```"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 2,
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "zR9ZbuQCH7FX",
+        "outputId": "6392c9ff-0863-4665-faf9-b3af9881c305"
+      },
+      "outputs": [
+        {
+          "output_type": "stream",
+          "name": "stdout",
+          "text": [
+            "\u001b[34m\u001b[1msegment/predict: \u001b[0mweights=['yolov5s-seg.pt'], source=data/images, data=data/coco128.yaml, imgsz=[640, 640], conf_thres=0.25, iou_thres=0.45, max_det=1000, device=, view_img=False, save_txt=False, save_conf=False, save_crop=False, nosave=False, classes=None, agnostic_nms=False, augment=False, visualize=False, update=False, project=runs/predict-seg, name=exp, exist_ok=False, line_thickness=3, hide_labels=False, hide_conf=False, half=False, dnn=False, vid_stride=1, retina_masks=False\n",
+            "YOLOv5 🚀 v6.2-257-g2ecaa96 Python-3.7.15 torch-1.12.1+cu113 CUDA:0 (Tesla T4, 15110MiB)\n",
+            "\n",
+            "Downloading https://github.com/ultralytics/yolov5/releases/download/v6.2/yolov5s-seg.pt to yolov5s-seg.pt...\n",
+            "100% 14.9M/14.9M [00:01<00:00, 9.09MB/s]\n",
+            "\n",
+            "Fusing layers... \n",
+            "YOLOv5s-seg summary: 224 layers, 7611485 parameters, 0 gradients, 26.4 GFLOPs\n",
+            "image 1/2 /content/yolov5/data/images/bus.jpg: 640x480 4 persons, 1 bus, 18.0ms\n",
+            "image 2/2 /content/yolov5/data/images/zidane.jpg: 384x640 2 persons, 1 tie, 13.5ms\n",
+            "Speed: 0.5ms pre-process, 15.7ms inference, 18.5ms NMS per image at shape (1, 3, 640, 640)\n",
+            "Results saved to \u001b[1mruns/predict-seg/exp\u001b[0m\n"
+          ]
+        }
+      ],
+      "source": [
+        "!python segment/predict.py --weights yolov5s-seg.pt --img 640 --conf 0.25 --source data/images\n",
+        "#display.Image(filename='runs/predict-seg/exp/zidane.jpg', width=600)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "hkAzDWJ7cWTr"
+      },
+      "source": [
+        "&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;\n",
+        "<img align=\"left\" src=\"https://user-images.githubusercontent.com/26833433/199030123-08c72f8d-6871-4116-8ed3-c373642cf28e.jpg\" width=\"600\">"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "0eq1SMWl6Sfn"
+      },
+      "source": [
+        "# 2. Validate\n",
+        "Validate a model's accuracy on the [COCO](https://cocodataset.org/#home) dataset's `val` or `test` splits. Models are downloaded automatically from the [latest YOLOv5 release](https://github.com/ultralytics/yolov5/releases). To show results by class use the `--verbose` flag."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 3,
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "WQPtK1QYVaD_",
+        "outputId": "4707734e-00c7-43da-d642-32c3c3fe3090"
+      },
+      "outputs": [
+        {
+          "output_type": "stream",
+          "name": "stdout",
+          "text": [
+            "Downloading https://github.com/ultralytics/yolov5/releases/download/v1.0/coco2017labels-segments.zip  ...\n",
+            "Downloading http://images.cocodataset.org/zips/val2017.zip ...\n",
+            "######################################################################## 100.0%\n",
+            "######################################################################## 100.0%\n"
+          ]
+        }
+      ],
+      "source": [
+        "# Download COCO val\n",
+        "!bash data/scripts/get_coco.sh --val --segments  # download (780M - 5000 images)"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 4,
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "X58w8JLpMnjH",
+        "outputId": "f96b700d-c779-4a34-930b-e85be4e58974"
+      },
+      "outputs": [
+        {
+          "output_type": "stream",
+          "name": "stdout",
+          "text": [
+            "\u001b[34m\u001b[1msegment/val: \u001b[0mdata=/content/yolov5/data/coco.yaml, weights=['yolov5s-seg.pt'], batch_size=32, imgsz=640, conf_thres=0.001, iou_thres=0.6, max_det=300, task=val, device=, workers=8, single_cls=False, augment=False, verbose=False, save_txt=False, save_hybrid=False, save_conf=False, save_json=False, project=runs/val-seg, name=exp, exist_ok=False, half=True, dnn=False\n",
+            "YOLOv5 🚀 v6.2-257-g2ecaa96 Python-3.7.15 torch-1.12.1+cu113 CUDA:0 (Tesla T4, 15110MiB)\n",
+            "\n",
+            "Fusing layers... \n",
+            "YOLOv5s-seg summary: 224 layers, 7611485 parameters, 0 gradients, 26.4 GFLOPs\n",
+            "\u001b[34m\u001b[1mval: \u001b[0mScanning /content/datasets/coco/val2017... 4952 images, 48 backgrounds, 0 corrupt: 100% 5000/5000 [00:03<00:00, 1409.04it/s]\n",
+            "\u001b[34m\u001b[1mval: \u001b[0mNew cache created: /content/datasets/coco/val2017.cache\n",
+            "                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95)     Mask(P          R      mAP50  mAP50-95): 100% 157/157 [01:53<00:00,  1.38it/s]\n",
+            "                   all       5000      36335      0.673      0.517      0.566      0.373      0.672       0.49      0.532      0.319\n",
+            "Speed: 0.8ms pre-process, 4.0ms inference, 2.8ms NMS per image at shape (32, 3, 640, 640)\n",
+            "Results saved to \u001b[1mruns/val-seg/exp\u001b[0m\n"
+          ]
+        }
+      ],
+      "source": [
+        "# Validate YOLOv5s-seg on COCO val\n",
+        "!python segment/val.py --weights yolov5s-seg.pt --data coco.yaml --img 640 --half"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "ZY2VXXXu74w5"
+      },
+      "source": [
+        "# 3. Train\n",
+        "\n",
+        "<p align=\"\"><a href=\"https://roboflow.com/?ref=ultralytics\"><img width=\"1000\" src=\"https://github.com/ultralytics/assets/raw/master/im/integrations-loop.png\"/></a></p>\n",
+        "Close the active learning loop by sampling images from your inference conditions with the `roboflow` pip package\n",
+        "<br><br>\n",
+        "\n",
+        "Train a YOLOv5s-seg model on the [COCO128](https://www.kaggle.com/ultralytics/coco128) dataset with `--data coco128-seg.yaml`, starting from pretrained `--weights yolov5s-seg.pt`, or from randomly initialized `--weights '' --cfg yolov5s-seg.yaml`.\n",
+        "\n",
+        "- **Pretrained [Models](https://github.com/ultralytics/yolov5/tree/master/models)** are downloaded\n",
+        "automatically from the [latest YOLOv5 release](https://github.com/ultralytics/yolov5/releases)\n",
+        "- **[Datasets](https://github.com/ultralytics/yolov5/tree/master/data)** available for autodownload include: [COCO](https://github.com/ultralytics/yolov5/blob/master/data/coco.yaml), [COCO128](https://github.com/ultralytics/yolov5/blob/master/data/coco128.yaml), [VOC](https://github.com/ultralytics/yolov5/blob/master/data/VOC.yaml), [Argoverse](https://github.com/ultralytics/yolov5/blob/master/data/Argoverse.yaml), [VisDrone](https://github.com/ultralytics/yolov5/blob/master/data/VisDrone.yaml), [GlobalWheat](https://github.com/ultralytics/yolov5/blob/master/data/GlobalWheat2020.yaml), [xView](https://github.com/ultralytics/yolov5/blob/master/data/xView.yaml), [Objects365](https://github.com/ultralytics/yolov5/blob/master/data/Objects365.yaml), [SKU-110K](https://github.com/ultralytics/yolov5/blob/master/data/SKU-110K.yaml).\n",
+        "- **Training Results** are saved to `runs/train-seg/` with incrementing run directories, i.e. `runs/train-seg/exp2`, `runs/train-seg/exp3` etc.\n",
+        "<br><br>\n",
+        "\n",
+        "A **Mosaic Dataloader** is used for training which combines 4 images into 1 mosaic.\n",
+        "\n",
+        "## Train on Custom Data with Roboflow 🌟 NEW\n",
+        "\n",
+        "[Roboflow](https://roboflow.com/?ref=ultralytics) enables you to easily **organize, label, and prepare** a high quality dataset with your own custom data. Roboflow also makes it easy to establish an active learning pipeline, collaborate with your team on dataset improvement, and integrate directly into your model building workflow with the `roboflow` pip package.\n",
+        "\n",
+        "- Custom Training Example: [https://blog.roboflow.com/train-yolov5-instance-segmentation-custom-dataset/](https://blog.roboflow.com/train-yolov5-instance-segmentation-custom-dataset/?ref=ultralytics)\n",
+        "- Custom Training Notebook: [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/drive/1JTz7kpmHsg-5qwVz2d2IH3AaenI1tv0N?usp=sharing)\n",
+        "<br>\n",
+        "\n",
+        "<p align=\"\"><a href=\"https://roboflow.com/?ref=ultralytics\"><img width=\"480\" src=\"https://robflow-public-assets.s3.amazonaws.com/how-to-train-yolov5-segmentation-annotation.gif\"/></a></p>Label images lightning fast (including with model-assisted labeling)"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "i3oKtE4g-aNn"
+      },
+      "outputs": [],
+      "source": [
+        "#@title Select YOLOv5 🚀 logger {run: 'auto'}\n",
+        "logger = 'TensorBoard' #@param ['TensorBoard', 'Comet', 'ClearML']\n",
+        "\n",
+        "if logger == 'TensorBoard':\n",
+        "  %load_ext tensorboard\n",
+        "  %tensorboard --logdir runs/train-seg\n",
+        "elif logger == 'Comet':\n",
+        "  %pip install -q comet_ml\n",
+        "  import comet_ml; comet_ml.init()\n",
+        "elif logger == 'ClearML':\n",
+        "  import clearml; clearml.browser_login()"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 5,
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "1NcFxRcFdJ_O",
+        "outputId": "2cdb19cc-69af-4c90-f8de-af02dfedba91"
+      },
+      "outputs": [
+        {
+          "output_type": "stream",
+          "name": "stdout",
+          "text": [
+            "\u001b[34m\u001b[1msegment/train: \u001b[0mweights=yolov5s-seg.pt, cfg=, data=coco128-seg.yaml, hyp=data/hyps/hyp.scratch-low.yaml, epochs=3, batch_size=16, imgsz=640, rect=False, resume=False, nosave=False, noval=False, noautoanchor=False, noplots=False, evolve=None, bucket=, cache=ram, image_weights=False, device=, multi_scale=False, single_cls=False, optimizer=SGD, sync_bn=False, workers=8, project=runs/train-seg, name=exp, exist_ok=False, quad=False, cos_lr=False, label_smoothing=0.0, patience=100, freeze=[0], save_period=-1, seed=0, local_rank=-1, mask_ratio=4, no_overlap=False\n",
+            "\u001b[34m\u001b[1mgithub: \u001b[0mup to date with https://github.com/ultralytics/yolov5 ✅\n",
+            "YOLOv5 🚀 v6.2-257-g2ecaa96 Python-3.7.15 torch-1.12.1+cu113 CUDA:0 (Tesla T4, 15110MiB)\n",
+            "\n",
+            "\u001b[34m\u001b[1mhyperparameters: \u001b[0mlr0=0.01, lrf=0.01, momentum=0.937, weight_decay=0.0005, warmup_epochs=3.0, warmup_momentum=0.8, warmup_bias_lr=0.1, box=0.05, cls=0.5, cls_pw=1.0, obj=1.0, obj_pw=1.0, iou_t=0.2, anchor_t=4.0, fl_gamma=0.0, hsv_h=0.015, hsv_s=0.7, hsv_v=0.4, degrees=0.0, translate=0.1, scale=0.5, shear=0.0, perspective=0.0, flipud=0.0, fliplr=0.5, mosaic=1.0, mixup=0.0, copy_paste=0.0\n",
+            "\u001b[34m\u001b[1mTensorBoard: \u001b[0mStart with 'tensorboard --logdir runs/train-seg', view at http://localhost:6006/\n",
+            "\n",
+            "Dataset not found ⚠️, missing paths ['/content/datasets/coco128-seg/images/train2017']\n",
+            "Downloading https://ultralytics.com/assets/coco128-seg.zip to coco128-seg.zip...\n",
+            "100% 6.79M/6.79M [00:01<00:00, 5.87MB/s]\n",
+            "Dataset download success ✅ (2.1s), saved to \u001b[1m/content/datasets\u001b[0m\n",
+            "\n",
+            "                 from  n    params  module                                  arguments                     \n",
+            "  0                -1  1      3520  models.common.Conv                      [3, 32, 6, 2, 2]              \n",
+            "  1                -1  1     18560  models.common.Conv                      [32, 64, 3, 2]                \n",
+            "  2                -1  1     18816  models.common.C3                        [64, 64, 1]                   \n",
+            "  3                -1  1     73984  models.common.Conv                      [64, 128, 3, 2]               \n",
+            "  4                -1  2    115712  models.common.C3                        [128, 128, 2]                 \n",
+            "  5                -1  1    295424  models.common.Conv                      [128, 256, 3, 2]              \n",
+            "  6                -1  3    625152  models.common.C3                        [256, 256, 3]                 \n",
+            "  7                -1  1   1180672  models.common.Conv                      [256, 512, 3, 2]              \n",
+            "  8                -1  1   1182720  models.common.C3                        [512, 512, 1]                 \n",
+            "  9                -1  1    656896  models.common.SPPF                      [512, 512, 5]                 \n",
+            " 10                -1  1    131584  models.common.Conv                      [512, 256, 1, 1]              \n",
+            " 11                -1  1         0  torch.nn.modules.upsampling.Upsample    [None, 2, 'nearest']          \n",
+            " 12           [-1, 6]  1         0  models.common.Concat                    [1]                           \n",
+            " 13                -1  1    361984  models.common.C3                        [512, 256, 1, False]          \n",
+            " 14                -1  1     33024  models.common.Conv                      [256, 128, 1, 1]              \n",
+            " 15                -1  1         0  torch.nn.modules.upsampling.Upsample    [None, 2, 'nearest']          \n",
+            " 16           [-1, 4]  1         0  models.common.Concat                    [1]                           \n",
+            " 17                -1  1     90880  models.common.C3                        [256, 128, 1, False]          \n",
+            " 18                -1  1    147712  models.common.Conv                      [128, 128, 3, 2]              \n",
+            " 19          [-1, 14]  1         0  models.common.Concat                    [1]                           \n",
+            " 20                -1  1    296448  models.common.C3                        [256, 256, 1, False]          \n",
+            " 21                -1  1    590336  models.common.Conv                      [256, 256, 3, 2]              \n",
+            " 22          [-1, 10]  1         0  models.common.Concat                    [1]                           \n",
+            " 23                -1  1   1182720  models.common.C3                        [512, 512, 1, False]          \n",
+            " 24      [17, 20, 23]  1    615133  models.yolo.Segment                     [80, [[10, 13, 16, 30, 33, 23], [30, 61, 62, 45, 59, 119], [116, 90, 156, 198, 373, 326]], 32, 128, [128, 256, 512]]\n",
+            "Model summary: 225 layers, 7621277 parameters, 7621277 gradients, 26.6 GFLOPs\n",
+            "\n",
+            "Transferred 367/367 items from yolov5s-seg.pt\n",
+            "\u001b[34m\u001b[1mAMP: \u001b[0mchecks passed ✅\n",
+            "\u001b[34m\u001b[1moptimizer:\u001b[0m SGD(lr=0.01) with parameter groups 60 weight(decay=0.0), 63 weight(decay=0.0005), 63 bias\n",
+            "\u001b[34m\u001b[1malbumentations: \u001b[0mBlur(p=0.01, blur_limit=(3, 7)), MedianBlur(p=0.01, blur_limit=(3, 7)), ToGray(p=0.01), CLAHE(p=0.01, clip_limit=(1, 4.0), tile_grid_size=(8, 8))\n",
+            "\u001b[34m\u001b[1mtrain: \u001b[0mScanning /content/datasets/coco128-seg/labels/train2017... 126 images, 2 backgrounds, 0 corrupt: 100% 128/128 [00:00<00:00, 1439.54it/s]\n",
+            "\u001b[34m\u001b[1mtrain: \u001b[0mNew cache created: /content/datasets/coco128-seg/labels/train2017.cache\n",
+            "\u001b[34m\u001b[1mtrain: \u001b[0mCaching images (0.1GB ram): 100% 128/128 [00:00<00:00, 253.53it/s]\n",
+            "\u001b[34m\u001b[1mval: \u001b[0mScanning /content/datasets/coco128-seg/labels/train2017.cache... 126 images, 2 backgrounds, 0 corrupt: 100% 128/128 [00:00<?, ?it/s]\n",
+            "\u001b[34m\u001b[1mval: \u001b[0mCaching images (0.1GB ram): 100% 128/128 [00:01<00:00, 93.82it/s]\n",
+            "\n",
+            "\u001b[34m\u001b[1mAutoAnchor: \u001b[0m4.27 anchors/target, 0.994 Best Possible Recall (BPR). Current anchors are a good fit to dataset ✅\n",
+            "Plotting labels to runs/train-seg/exp/labels.jpg... \n",
+            "Image sizes 640 train, 640 val\n",
+            "Using 2 dataloader workers\n",
+            "Logging results to \u001b[1mruns/train-seg/exp\u001b[0m\n",
+            "Starting training for 3 epochs...\n",
+            "\n",
+            "      Epoch    GPU_mem   box_loss   seg_loss   obj_loss   cls_loss  Instances       Size\n",
+            "        0/2      4.92G     0.0417    0.04646    0.06066    0.02126        192        640: 100% 8/8 [00:07<00:00,  1.11it/s]\n",
+            "                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95)     Mask(P          R      mAP50  mAP50-95): 100% 4/4 [00:02<00:00,  1.85it/s]\n",
+            "                   all        128        929      0.737      0.649      0.715      0.492      0.719      0.617      0.658      0.408\n",
+            "\n",
+            "      Epoch    GPU_mem   box_loss   seg_loss   obj_loss   cls_loss  Instances       Size\n",
+            "        1/2      6.29G    0.04157    0.04503    0.05772    0.01777        208        640: 100% 8/8 [00:09<00:00,  1.19s/it]\n",
+            "                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95)     Mask(P          R      mAP50  mAP50-95): 100% 4/4 [00:02<00:00,  1.75it/s]\n",
+            "                   all        128        929      0.756      0.674      0.738      0.506      0.725       0.64       0.68      0.422\n",
+            "\n",
+            "      Epoch    GPU_mem   box_loss   seg_loss   obj_loss   cls_loss  Instances       Size\n",
+            "        2/2      6.29G     0.0425    0.04793    0.06784    0.01863        161        640: 100% 8/8 [00:04<00:00,  1.99it/s]\n",
+            "                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95)     Mask(P          R      mAP50  mAP50-95): 100% 4/4 [00:02<00:00,  1.81it/s]\n",
+            "                   all        128        929      0.736      0.694      0.747      0.522      0.769      0.622      0.683      0.427\n",
+            "\n",
+            "3 epochs completed in 0.008 hours.\n",
+            "Optimizer stripped from runs/train-seg/exp/weights/last.pt, 15.6MB\n",
+            "Optimizer stripped from runs/train-seg/exp/weights/best.pt, 15.6MB\n",
+            "\n",
+            "Validating runs/train-seg/exp/weights/best.pt...\n",
+            "Fusing layers... \n",
+            "Model summary: 165 layers, 7611485 parameters, 0 gradients, 26.4 GFLOPs\n",
+            "                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95)     Mask(P          R      mAP50  mAP50-95): 100% 4/4 [00:06<00:00,  1.58s/it]\n",
+            "                   all        128        929      0.738      0.694      0.746      0.522      0.759      0.625      0.682      0.426\n",
+            "                person        128        254      0.845      0.756      0.836       0.55      0.861      0.669      0.759      0.407\n",
+            "               bicycle        128          6      0.475      0.333      0.549      0.341      0.711      0.333      0.526      0.322\n",
+            "                   car        128         46      0.612      0.565      0.539      0.257      0.555      0.435      0.477      0.171\n",
+            "            motorcycle        128          5       0.73        0.8      0.752      0.571      0.747        0.8      0.752       0.42\n",
+            "              airplane        128          6          1      0.943      0.995      0.732       0.92      0.833      0.839      0.555\n",
+            "                   bus        128          7      0.677      0.714      0.722      0.653      0.711      0.714      0.722      0.593\n",
+            "                 train        128          3          1      0.951      0.995      0.551          1      0.884      0.995      0.781\n",
+            "                 truck        128         12      0.555      0.417      0.457      0.285      0.624      0.417      0.397      0.277\n",
+            "                  boat        128          6      0.624        0.5      0.584      0.186          1      0.326      0.412      0.133\n",
+            "         traffic light        128         14      0.513      0.302      0.411      0.247      0.435      0.214      0.376      0.251\n",
+            "             stop sign        128          2      0.824          1      0.995      0.796      0.906          1      0.995      0.747\n",
+            "                 bench        128          9       0.75      0.667      0.763      0.367      0.724      0.585      0.698      0.209\n",
+            "                  bird        128         16      0.961          1      0.995      0.686      0.918      0.938       0.91      0.525\n",
+            "                   cat        128          4      0.771      0.857      0.945      0.752       0.76        0.8      0.945      0.728\n",
+            "                   dog        128          9      0.987      0.778      0.963      0.681          1      0.705       0.89      0.574\n",
+            "                 horse        128          2      0.703          1      0.995      0.697      0.759          1      0.995      0.249\n",
+            "              elephant        128         17      0.916      0.882       0.93      0.691      0.811      0.765      0.829      0.537\n",
+            "                  bear        128          1      0.664          1      0.995      0.995      0.701          1      0.995      0.895\n",
+            "                 zebra        128          4      0.864          1      0.995      0.921      0.879          1      0.995      0.804\n",
+            "               giraffe        128          9      0.883      0.889       0.94      0.683      0.845      0.778       0.78      0.463\n",
+            "              backpack        128          6          1       0.59      0.701      0.372          1      0.474       0.52      0.252\n",
+            "              umbrella        128         18      0.654      0.839      0.887       0.52      0.517      0.556      0.427      0.229\n",
+            "               handbag        128         19       0.54      0.211      0.408      0.221      0.796      0.206      0.396      0.196\n",
+            "                   tie        128          7      0.864      0.857      0.857      0.577      0.925      0.857      0.857      0.534\n",
+            "              suitcase        128          4      0.716          1      0.945      0.647      0.767          1      0.945      0.634\n",
+            "               frisbee        128          5      0.708        0.8      0.761      0.643      0.737        0.8      0.761      0.501\n",
+            "                  skis        128          1      0.691          1      0.995      0.796      0.761          1      0.995      0.199\n",
+            "             snowboard        128          7      0.918      0.857      0.904      0.604       0.32      0.286      0.235      0.137\n",
+            "           sports ball        128          6      0.902      0.667      0.701      0.466      0.727        0.5      0.497      0.471\n",
+            "                  kite        128         10      0.586        0.4      0.511      0.231      0.663      0.394      0.417      0.139\n",
+            "          baseball bat        128          4      0.359        0.5      0.401      0.169      0.631        0.5      0.526      0.133\n",
+            "        baseball glove        128          7          1      0.519       0.58      0.327      0.687      0.286      0.455      0.328\n",
+            "            skateboard        128          5      0.729        0.8      0.862      0.631      0.599        0.6      0.604      0.379\n",
+            "         tennis racket        128          7       0.57      0.714      0.645      0.448      0.608      0.714      0.645      0.412\n",
+            "                bottle        128         18      0.469      0.393      0.537      0.357      0.661      0.389      0.543      0.349\n",
+            "            wine glass        128         16      0.677      0.938      0.866      0.441       0.53      0.625       0.67      0.334\n",
+            "                   cup        128         36      0.777      0.722      0.812      0.466      0.725      0.583      0.762      0.467\n",
+            "                  fork        128          6      0.948      0.333      0.425       0.27      0.527      0.167       0.18      0.102\n",
+            "                 knife        128         16      0.757      0.587      0.669      0.458       0.79        0.5      0.552       0.34\n",
+            "                 spoon        128         22       0.74      0.364      0.559      0.269      0.925      0.364      0.513      0.213\n",
+            "                  bowl        128         28      0.766      0.714      0.725      0.559      0.803      0.584      0.665      0.353\n",
+            "                banana        128          1      0.408          1      0.995      0.398      0.539          1      0.995      0.497\n",
+            "              sandwich        128          2          1          0      0.695      0.536          1          0      0.498      0.448\n",
+            "                orange        128          4      0.467          1      0.995      0.693      0.518          1      0.995      0.663\n",
+            "              broccoli        128         11      0.462      0.455      0.383      0.259      0.548      0.455      0.384      0.256\n",
+            "                carrot        128         24      0.631      0.875       0.77      0.533      0.757      0.909      0.853      0.499\n",
+            "               hot dog        128          2      0.555          1      0.995      0.995      0.578          1      0.995      0.796\n",
+            "                 pizza        128          5       0.89        0.8      0.962      0.796          1      0.778      0.962      0.766\n",
+            "                 donut        128         14      0.695          1      0.893      0.772      0.704          1      0.893      0.696\n",
+            "                  cake        128          4      0.826          1      0.995       0.92      0.862          1      0.995      0.846\n",
+            "                 chair        128         35       0.53      0.571      0.613      0.336       0.67        0.6      0.538      0.271\n",
+            "                 couch        128          6      0.972      0.667      0.833      0.627          1       0.62      0.696      0.394\n",
+            "          potted plant        128         14        0.7      0.857      0.883      0.552      0.836      0.857      0.883      0.473\n",
+            "                   bed        128          3      0.979      0.667       0.83      0.366          1          0       0.83      0.373\n",
+            "          dining table        128         13      0.775      0.308      0.505      0.364      0.644      0.231       0.25     0.0804\n",
+            "                toilet        128          2      0.836          1      0.995      0.846      0.887          1      0.995      0.797\n",
+            "                    tv        128          2        0.6          1      0.995      0.846      0.655          1      0.995      0.896\n",
+            "                laptop        128          3      0.822      0.333      0.445      0.307          1          0      0.392       0.12\n",
+            "                 mouse        128          2          1          0          0          0          1          0          0          0\n",
+            "                remote        128          8      0.745        0.5       0.62      0.459      0.821        0.5      0.624      0.449\n",
+            "            cell phone        128          8      0.686      0.375      0.502      0.272      0.488       0.25       0.28      0.132\n",
+            "             microwave        128          3      0.831          1      0.995      0.722      0.867          1      0.995      0.592\n",
+            "                  oven        128          5      0.439        0.4      0.435      0.294      0.823        0.6      0.645      0.418\n",
+            "                  sink        128          6      0.677        0.5      0.565      0.448      0.722        0.5       0.46      0.362\n",
+            "          refrigerator        128          5      0.533        0.8      0.783      0.524      0.558        0.8      0.783      0.527\n",
+            "                  book        128         29      0.732      0.379      0.423      0.196       0.69      0.207       0.38      0.131\n",
+            "                 clock        128          9      0.889      0.778      0.917      0.677      0.908      0.778      0.875      0.604\n",
+            "                  vase        128          2      0.375          1      0.995      0.995      0.455          1      0.995      0.796\n",
+            "              scissors        128          1          1          0     0.0166    0.00166          1          0          0          0\n",
+            "            teddy bear        128         21      0.813      0.829      0.841      0.457      0.826      0.678      0.786      0.422\n",
+            "            toothbrush        128          5      0.806          1      0.995      0.733      0.991          1      0.995      0.628\n",
+            "Results saved to \u001b[1mruns/train-seg/exp\u001b[0m\n"
+          ]
+        }
+      ],
+      "source": [
+        "# Train YOLOv5s on COCO128 for 3 epochs\n",
+        "!python segment/train.py --img 640 --batch 16 --epochs 3 --data coco128-seg.yaml --weights yolov5s-seg.pt --cache"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "15glLzbQx5u0"
+      },
+      "source": [
+        "# 4. Visualize"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "nWOsI5wJR1o3"
+      },
+      "source": [
+        "## Comet Logging and Visualization 🌟 NEW\n",
+        "[Comet](https://bit.ly/yolov5-readme-comet) is now fully integrated with YOLOv5. Track and visualize model metrics in real time, save your hyperparameters, datasets, and model checkpoints, and visualize your model predictions with [Comet Custom Panels](https://bit.ly/yolov5-colab-comet-panels)! Comet makes sure you never lose track of your work and makes it easy to share results and collaborate across teams of all sizes! \n",
+        "\n",
+        "Getting started is easy:\n",
+        "```shell\n",
+        "pip install comet_ml  # 1. install\n",
+        "export COMET_API_KEY=<Your API Key>  # 2. paste API key\n",
+        "python train.py --img 640 --epochs 3 --data coco128.yaml --weights yolov5s.pt  # 3. train\n",
+        "```\n",
+        "\n",
+        "To learn more about all of the supported Comet features for this integration, check out the [Comet Tutorial](https://github.com/ultralytics/yolov5/tree/master/utils/loggers/comet). If you'd like to learn more about Comet, head over to our [documentation](https://bit.ly/yolov5-colab-comet-docs). Get started by trying out the Comet Colab Notebook:\n",
+        "[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/drive/1RG0WOQyxlDlo5Km8GogJpIEJlg_5lyYO?usp=sharing)\n",
+        "\n",
+        "<img width=\"1920\" alt=\"yolo-ui\" src=\"https://user-images.githubusercontent.com/26833433/202851203-164e94e1-2238-46dd-91f8-de020e9d6b41.png\">"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "Lay2WsTjNJzP"
+      },
+      "source": [
+        "## ClearML Logging and Automation 🌟 NEW\n",
+        "\n",
+        "[ClearML](https://cutt.ly/yolov5-notebook-clearml) is completely integrated into YOLOv5 to track your experimentation, manage dataset versions and even remotely execute training runs. To enable ClearML (check cells above):\n",
+        "\n",
+        "- `pip install clearml`\n",
+        "- run `clearml-init` to connect to a ClearML server (**deploy your own [open-source server](https://github.com/allegroai/clearml-server)**, or use our [free hosted server](https://cutt.ly/yolov5-notebook-clearml))\n",
+        "\n",
+        "You'll get all the great expected features from an experiment manager: live updates, model upload, experiment comparison etc. but ClearML also tracks uncommitted changes and installed packages for example. Thanks to that ClearML Tasks (which is what we call experiments) are also reproducible on different machines! With only 1 extra line, we can schedule a YOLOv5 training task on a queue to be executed by any number of ClearML Agents (workers).\n",
+        "\n",
+        "You can use ClearML Data to version your dataset and then pass it to YOLOv5 simply using its unique ID. This will help you keep track of your data without adding extra hassle. Explore the [ClearML Tutorial](https://github.com/ultralytics/yolov5/tree/master/utils/loggers/clearml) for details!\n",
+        "\n",
+        "<a href=\"https://cutt.ly/yolov5-notebook-clearml\">\n",
+        "<img alt=\"ClearML Experiment Management UI\" src=\"https://github.com/thepycoder/clearml_screenshots/raw/main/scalars.jpg\" width=\"1280\"/></a>"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "-WPvRbS5Swl6"
+      },
+      "source": [
+        "## Local Logging\n",
+        "\n",
+        "Training results are automatically logged with [Tensorboard](https://www.tensorflow.org/tensorboard) and [CSV](https://github.com/ultralytics/yolov5/pull/4148) loggers to `runs/train`, with a new experiment directory created for each new training as `runs/train/exp2`, `runs/train/exp3`, etc.\n",
+        "\n",
+        "This directory contains train and val statistics, mosaics, labels, predictions and augmentated mosaics, as well as metrics and charts including precision-recall (PR) curves and confusion matrices. \n",
+        "\n",
+        "<img alt=\"Local logging results\" src=\"https://user-images.githubusercontent.com/26833433/183222430-e1abd1b7-782c-4cde-b04d-ad52926bf818.jpg\" width=\"1280\"/>\n"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "Zelyeqbyt3GD"
+      },
+      "source": [
+        "# Environments\n",
+        "\n",
+        "YOLOv5 may be run in any of the following up-to-date verified environments (with all dependencies including [CUDA](https://developer.nvidia.com/cuda)/[CUDNN](https://developer.nvidia.com/cudnn), [Python](https://www.python.org/) and [PyTorch](https://pytorch.org/) preinstalled):\n",
+        "\n",
+        "- **Notebooks** with free GPU: <a href=\"https://bit.ly/yolov5-paperspace-notebook\"><img src=\"https://assets.paperspace.io/img/gradient-badge.svg\" alt=\"Run on Gradient\"></a> <a href=\"https://colab.research.google.com/github/ultralytics/yolov5/blob/master/tutorial.ipynb\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"></a> <a href=\"https://www.kaggle.com/ultralytics/yolov5\"><img src=\"https://kaggle.com/static/images/open-in-kaggle.svg\" alt=\"Open In Kaggle\"></a>\n",
+        "- **Google Cloud** Deep Learning VM. See [GCP Quickstart Guide](https://github.com/ultralytics/yolov5/wiki/GCP-Quickstart)\n",
+        "- **Amazon** Deep Learning AMI. See [AWS Quickstart Guide](https://github.com/ultralytics/yolov5/wiki/AWS-Quickstart)\n",
+        "- **Docker Image**. See [Docker Quickstart Guide](https://github.com/ultralytics/yolov5/wiki/Docker-Quickstart) <a href=\"https://hub.docker.com/r/ultralytics/yolov5\"><img src=\"https://img.shields.io/docker/pulls/ultralytics/yolov5?logo=docker\" alt=\"Docker Pulls\"></a>\n"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "6Qu7Iesl0p54"
+      },
+      "source": [
+        "# Status\n",
+        "\n",
+        "![YOLOv5 CI](https://github.com/ultralytics/yolov5/actions/workflows/ci-testing.yml/badge.svg)\n",
+        "\n",
+        "If this badge is green, all [YOLOv5 GitHub Actions](https://github.com/ultralytics/yolov5/actions) Continuous Integration (CI) tests are currently passing. CI tests verify correct operation of YOLOv5 training ([train.py](https://github.com/ultralytics/yolov5/blob/master/train.py)), testing ([val.py](https://github.com/ultralytics/yolov5/blob/master/val.py)), inference ([detect.py](https://github.com/ultralytics/yolov5/blob/master/detect.py)) and export ([export.py](https://github.com/ultralytics/yolov5/blob/master/export.py)) on macOS, Windows, and Ubuntu every 24 hours and on every commit.\n"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "IEijrePND_2I"
+      },
+      "source": [
+        "# Appendix\n",
+        "\n",
+        "Additional content below."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "GMusP4OAxFu6"
+      },
+      "outputs": [],
+      "source": [
+        "# YOLOv5 PyTorch HUB Inference (DetectionModels only)\n",
+        "import torch\n",
+        "\n",
+        "model = torch.hub.load('ultralytics/yolov5', 'yolov5s-seg')  # yolov5n - yolov5x6 or custom\n",
+        "im = 'https://ultralytics.com/images/zidane.jpg'  # file, Path, PIL.Image, OpenCV, nparray, list\n",
+        "results = model(im)  # inference\n",
+        "results.print()  # or .show(), .save(), .crop(), .pandas(), etc."
+      ]
+    }
+  ],
+  "metadata": {
+    "accelerator": "GPU",
+    "colab": {
+      "name": "YOLOv5 Segmentation Tutorial",
+      "provenance": [],
+      "toc_visible": true
+    },
+    "kernelspec": {
+      "display_name": "Python 3 (ipykernel)",
+      "language": "python",
+      "name": "python3"
+    },
+    "language_info": {
+      "codemirror_mode": {
+        "name": "ipython",
+        "version": 3
+      },
+      "file_extension": ".py",
+      "mimetype": "text/x-python",
+      "name": "python",
+      "nbconvert_exporter": "python",
+      "pygments_lexer": "ipython3",
+      "version": "3.7.12"
+    }
+  },
+  "nbformat": 4,
+  "nbformat_minor": 0
+}

segment/val.py

+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+"""
+Validate a trained YOLOv5 segment model on a segment dataset
+
+Usage:
+    $ bash data/scripts/get_coco.sh --val --segments  # download COCO-segments val split (1G, 5000 images)
+    $ python segment/val.py --weights yolov5s-seg.pt --data coco.yaml --img 640  # validate COCO-segments
+
+Usage - formats:
+    $ python segment/val.py --weights yolov5s-seg.pt                 # PyTorch
+                                      yolov5s-seg.torchscript        # TorchScript
+                                      yolov5s-seg.onnx               # ONNX Runtime or OpenCV DNN with --dnn
+                                      yolov5s-seg_openvino_label     # OpenVINO
+                                      yolov5s-seg.engine             # TensorRT
+                                      yolov5s-seg.mlmodel            # CoreML (macOS-only)
+                                      yolov5s-seg_saved_model        # TensorFlow SavedModel
+                                      yolov5s-seg.pb                 # TensorFlow GraphDef
+                                      yolov5s-seg.tflite             # TensorFlow Lite
+                                      yolov5s-seg_edgetpu.tflite     # TensorFlow Edge TPU
+                                      yolov5s-seg_paddle_model       # PaddlePaddle
+"""
+
+import argparse
+import json
+import os
+import sys
+from multiprocessing.pool import ThreadPool
+from pathlib import Path
+
+import numpy as np
+import torch
+from tqdm import tqdm
+
+FILE = Path(__file__).resolve()
+ROOT = FILE.parents[1]  # YOLOv5 root directory
+if str(ROOT) not in sys.path:
+    sys.path.append(str(ROOT))  # add ROOT to PATH
+ROOT = Path(os.path.relpath(ROOT, Path.cwd()))  # relative
+
+import torch.nn.functional as F
+
+from models.common import DetectMultiBackend
+from models.yolo import SegmentationModel
+from utils.callbacks import Callbacks
+from utils.general import (LOGGER, NUM_THREADS, TQDM_BAR_FORMAT, Profile, check_dataset, check_img_size,
+                           check_requirements, check_yaml, coco80_to_coco91_class, colorstr, increment_path,
+                           non_max_suppression, print_args, scale_boxes, xywh2xyxy, xyxy2xywh)
+from utils.metrics import ConfusionMatrix, box_iou
+from utils.plots import output_to_target, plot_val_study
+from utils.segment.dataloaders import create_dataloader
+from utils.segment.general import mask_iou, process_mask, process_mask_upsample, scale_image
+from utils.segment.metrics import Metrics, ap_per_class_box_and_mask
+from utils.segment.plots import plot_images_and_masks
+from utils.torch_utils import de_parallel, select_device, smart_inference_mode
+
+
+def save_one_txt(predn, save_conf, shape, file):
+    # Save one txt result
+    gn = torch.tensor(shape)[[1, 0, 1, 0]]  # normalization gain whwh
+    for *xyxy, conf, cls in predn.tolist():
+        xywh = (xyxy2xywh(torch.tensor(xyxy).view(1, 4)) / gn).view(-1).tolist()  # normalized xywh
+        line = (cls, *xywh, conf) if save_conf else (cls, *xywh)  # label format
+        with open(file, 'a') as f:
+            f.write(('%g ' * len(line)).rstrip() % line + '\n')
+
+
+def save_one_json(predn, jdict, path, class_map, pred_masks):
+    # Save one JSON result {"image_id": 42, "category_id": 18, "bbox": [258.15, 41.29, 348.26, 243.78], "score": 0.236}
+    from pycocotools.mask import encode
+
+    def single_encode(x):
+        rle = encode(np.asarray(x[:, :, None], order="F", dtype="uint8"))[0]
+        rle["counts"] = rle["counts"].decode("utf-8")
+        return rle
+
+    image_id = int(path.stem) if path.stem.isnumeric() else path.stem
+    box = xyxy2xywh(predn[:, :4])  # xywh
+    box[:, :2] -= box[:, 2:] / 2  # xy center to top-left corner
+    pred_masks = np.transpose(pred_masks, (2, 0, 1))
+    with ThreadPool(NUM_THREADS) as pool:
+        rles = pool.map(single_encode, pred_masks)
+    for i, (p, b) in enumerate(zip(predn.tolist(), box.tolist())):
+        jdict.append({
+            'image_id': image_id,
+            'category_id': class_map[int(p[5])],
+            'bbox': [round(x, 3) for x in b],
+            'score': round(p[4], 5),
+            'segmentation': rles[i]})
+
+
+def process_batch(detections, labels, iouv, pred_masks=None, gt_masks=None, overlap=False, masks=False):
+    """
+    Return correct prediction matrix
+    Arguments:
+        detections (array[N, 6]), x1, y1, x2, y2, conf, class
+        labels (array[M, 5]), class, x1, y1, x2, y2
+    Returns:
+        correct (array[N, 10]), for 10 IoU levels
+    """
+    if masks:
+        if overlap:
+            nl = len(labels)
+            index = torch.arange(nl, device=gt_masks.device).view(nl, 1, 1) + 1
+            gt_masks = gt_masks.repeat(nl, 1, 1)  # shape(1,640,640) -> (n,640,640)
+            gt_masks = torch.where(gt_masks == index, 1.0, 0.0)
+        if gt_masks.shape[1:] != pred_masks.shape[1:]:
+            gt_masks = F.interpolate(gt_masks[None], pred_masks.shape[1:], mode="bilinear", align_corners=False)[0]
+            gt_masks = gt_masks.gt_(0.5)
+        iou = mask_iou(gt_masks.view(gt_masks.shape[0], -1), pred_masks.view(pred_masks.shape[0], -1))
+    else:  # boxes
+        iou = box_iou(labels[:, 1:], detections[:, :4])
+
+    correct = np.zeros((detections.shape[0], iouv.shape[0])).astype(bool)
+    correct_class = labels[:, 0:1] == detections[:, 5]
+    for i in range(len(iouv)):
+        x = torch.where((iou >= iouv[i]) & correct_class)  # IoU > threshold and classes match
+        if x[0].shape[0]:
+            matches = torch.cat((torch.stack(x, 1), iou[x[0], x[1]][:, None]), 1).cpu().numpy()  # [label, detect, iou]
+            if x[0].shape[0] > 1:
+                matches = matches[matches[:, 2].argsort()[::-1]]
+                matches = matches[np.unique(matches[:, 1], return_index=True)[1]]
+                # matches = matches[matches[:, 2].argsort()[::-1]]
+                matches = matches[np.unique(matches[:, 0], return_index=True)[1]]
+            correct[matches[:, 1].astype(int), i] = True
+    return torch.tensor(correct, dtype=torch.bool, device=iouv.device)
+
+
+@smart_inference_mode()
+def run(
+        data,
+        weights=None,  # model.pt path(s)
+        batch_size=32,  # batch size
+        imgsz=640,  # inference size (pixels)
+        conf_thres=0.001,  # confidence threshold
+        iou_thres=0.6,  # NMS IoU threshold
+        max_det=300,  # maximum detections per image
+        task='val',  # train, val, test, speed or study
+        device='',  # cuda device, i.e. 0 or 0,1,2,3 or cpu
+        workers=8,  # max dataloader workers (per RANK in DDP mode)
+        single_cls=False,  # treat as single-class dataset
+        augment=False,  # augmented inference
+        verbose=False,  # verbose output
+        save_txt=False,  # save results to *.txt
+        save_hybrid=False,  # save label+prediction hybrid results to *.txt
+        save_conf=False,  # save confidences in --save-txt labels
+        save_json=False,  # save a COCO-JSON results file
+        project=ROOT / 'runs/val-seg',  # save to project/name
+        name='exp',  # save to project/name
+        exist_ok=False,  # existing project/name ok, do not increment
+        half=True,  # use FP16 half-precision inference
+        dnn=False,  # use OpenCV DNN for ONNX inference
+        model=None,
+        dataloader=None,
+        save_dir=Path(''),
+        plots=True,
+        overlap=False,
+        mask_downsample_ratio=1,
+        compute_loss=None,
+        callbacks=Callbacks(),
+):
+    if save_json:
+        check_requirements(['pycocotools'])
+        process = process_mask_upsample  # more accurate
+    else:
+        process = process_mask  # faster
+
+    # Initialize/load model and set device
+    training = model is not None
+    if training:  # called by train.py
+        device, pt, jit, engine = next(model.parameters()).device, True, False, False  # get model device, PyTorch model
+        half &= device.type != 'cpu'  # half precision only supported on CUDA
+        model.half() if half else model.float()
+        nm = de_parallel(model).model[-1].nm  # number of masks
+    else:  # called directly
+        device = select_device(device, batch_size=batch_size)
+
+        # Directories
+        save_dir = increment_path(Path(project) / name, exist_ok=exist_ok)  # increment run
+        (save_dir / 'labels' if save_txt else save_dir).mkdir(parents=True, exist_ok=True)  # make dir
+
+        # Load model
+        model = DetectMultiBackend(weights, device=device, dnn=dnn, data=data, fp16=half)
+        stride, pt, jit, engine = model.stride, model.pt, model.jit, model.engine
+        imgsz = check_img_size(imgsz, s=stride)  # check image size
+        half = model.fp16  # FP16 supported on limited backends with CUDA
+        nm = de_parallel(model).model.model[-1].nm if isinstance(model, SegmentationModel) else 32  # number of masks
+        if engine:
+            batch_size = model.batch_size
+        else:
+            device = model.device
+            if not (pt or jit):
+                batch_size = 1  # export.py models default to batch-size 1
+                LOGGER.info(f'Forcing --batch-size 1 square inference (1,3,{imgsz},{imgsz}) for non-PyTorch models')
+
+        # Data
+        data = check_dataset(data)  # check
+
+    # Configure
+    model.eval()
+    cuda = device.type != 'cpu'
+    is_coco = isinstance(data.get('val'), str) and data['val'].endswith(f'coco{os.sep}val2017.txt')  # COCO dataset
+    nc = 1 if single_cls else int(data['nc'])  # number of classes
+    iouv = torch.linspace(0.5, 0.95, 10, device=device)  # iou vector for mAP@0.5:0.95
+    niou = iouv.numel()
+
+    # Dataloader
+    if not training:
+        if pt and not single_cls:  # check --weights are trained on --data
+            ncm = model.model.nc
+            assert ncm == nc, f'{weights} ({ncm} classes) trained on different --data than what you passed ({nc} ' \
+                              f'classes). Pass correct combination of --weights and --data that are trained together.'
+        model.warmup(imgsz=(1 if pt else batch_size, 3, imgsz, imgsz))  # warmup
+        pad, rect = (0.0, False) if task == 'speed' else (0.5, pt)  # square inference for benchmarks
+        task = task if task in ('train', 'val', 'test') else 'val'  # path to train/val/test images
+        dataloader = create_dataloader(data[task],
+                                       imgsz,
+                                       batch_size,
+                                       stride,
+                                       single_cls,
+                                       pad=pad,
+                                       rect=rect,
+                                       workers=workers,
+                                       prefix=colorstr(f'{task}: '),
+                                       overlap_mask=overlap,
+                                       mask_downsample_ratio=mask_downsample_ratio)[0]
+
+    seen = 0
+    confusion_matrix = ConfusionMatrix(nc=nc)
+    names = model.names if hasattr(model, 'names') else model.module.names  # get class names
+    if isinstance(names, (list, tuple)):  # old format
+        names = dict(enumerate(names))
+    class_map = coco80_to_coco91_class() if is_coco else list(range(1000))
+    s = ('%22s' + '%11s' * 10) % ('Class', 'Images', 'Instances', 'Box(P', "R", "mAP50", "mAP50-95)", "Mask(P", "R",
+                                  "mAP50", "mAP50-95)")
+    dt = Profile(), Profile(), Profile()
+    metrics = Metrics()
+    loss = torch.zeros(4, device=device)
+    jdict, stats = [], []
+    # callbacks.run('on_val_start')
+    pbar = tqdm(dataloader, desc=s, bar_format=TQDM_BAR_FORMAT)  # progress bar
+    for batch_i, (im, targets, paths, shapes, masks) in enumerate(pbar):
+        # callbacks.run('on_val_batch_start')
+        with dt[0]:
+            if cuda:
+                im = im.to(device, non_blocking=True)
+                targets = targets.to(device)
+                masks = masks.to(device)
+            masks = masks.float()
+            im = im.half() if half else im.float()  # uint8 to fp16/32
+            im /= 255  # 0 - 255 to 0.0 - 1.0
+            nb, _, height, width = im.shape  # batch size, channels, height, width
+
+        # Inference
+        with dt[1]:
+            preds, protos, train_out = model(im) if compute_loss else (*model(im, augment=augment)[:2], None)
+
+        # Loss
+        if compute_loss:
+            loss += compute_loss((train_out, protos), targets, masks)[1]  # box, obj, cls
+
+        # NMS
+        targets[:, 2:] *= torch.tensor((width, height, width, height), device=device)  # to pixels
+        lb = [targets[targets[:, 0] == i, 1:] for i in range(nb)] if save_hybrid else []  # for autolabelling
+        with dt[2]:
+            preds = non_max_suppression(preds,
+                                        conf_thres,
+                                        iou_thres,
+                                        labels=lb,
+                                        multi_label=True,
+                                        agnostic=single_cls,
+                                        max_det=max_det,
+                                        nm=nm)
+
+        # Metrics
+        plot_masks = []  # masks for plotting
+        for si, (pred, proto) in enumerate(zip(preds, protos)):
+            labels = targets[targets[:, 0] == si, 1:]
+            nl, npr = labels.shape[0], pred.shape[0]  # number of labels, predictions
+            path, shape = Path(paths[si]), shapes[si][0]
+            correct_masks = torch.zeros(npr, niou, dtype=torch.bool, device=device)  # init
+            correct_bboxes = torch.zeros(npr, niou, dtype=torch.bool, device=device)  # init
+            seen += 1
+
+            if npr == 0:
+                if nl:
+                    stats.append((correct_masks, correct_bboxes, *torch.zeros((2, 0), device=device), labels[:, 0]))
+                    if plots:
+                        confusion_matrix.process_batch(detections=None, labels=labels[:, 0])
+                continue
+
+            # Masks
+            midx = [si] if overlap else targets[:, 0] == si
+            gt_masks = masks[midx]
+            pred_masks = process(proto, pred[:, 6:], pred[:, :4], shape=im[si].shape[1:])
+
+            # Predictions
+            if single_cls:
+                pred[:, 5] = 0
+            predn = pred.clone()
+            scale_boxes(im[si].shape[1:], predn[:, :4], shape, shapes[si][1])  # native-space pred
+
+            # Evaluate
+            if nl:
+                tbox = xywh2xyxy(labels[:, 1:5])  # target boxes
+                scale_boxes(im[si].shape[1:], tbox, shape, shapes[si][1])  # native-space labels
+                labelsn = torch.cat((labels[:, 0:1], tbox), 1)  # native-space labels
+                correct_bboxes = process_batch(predn, labelsn, iouv)
+                correct_masks = process_batch(predn, labelsn, iouv, pred_masks, gt_masks, overlap=overlap, masks=True)
+                if plots:
+                    confusion_matrix.process_batch(predn, labelsn)
+            stats.append((correct_masks, correct_bboxes, pred[:, 4], pred[:, 5], labels[:, 0]))  # (conf, pcls, tcls)
+
+            pred_masks = torch.as_tensor(pred_masks, dtype=torch.uint8)
+            if plots and batch_i < 3:
+                plot_masks.append(pred_masks[:15].cpu())  # filter top 15 to plot
+
+            # Save/log
+            if save_txt:
+                save_one_txt(predn, save_conf, shape, file=save_dir / 'labels' / f'{path.stem}.txt')
+            if save_json:
+                pred_masks = scale_image(im[si].shape[1:],
+                                         pred_masks.permute(1, 2, 0).contiguous().cpu().numpy(), shape, shapes[si][1])
+                save_one_json(predn, jdict, path, class_map, pred_masks)  # append to COCO-JSON dictionary
+            # callbacks.run('on_val_image_end', pred, predn, path, names, im[si])
+
+        # Plot images
+        if plots and batch_i < 3:
+            if len(plot_masks):
+                plot_masks = torch.cat(plot_masks, dim=0)
+            plot_images_and_masks(im, targets, masks, paths, save_dir / f'val_batch{batch_i}_labels.jpg', names)
+            plot_images_and_masks(im, output_to_target(preds, max_det=15), plot_masks, paths,
+                                  save_dir / f'val_batch{batch_i}_pred.jpg', names)  # pred
+
+        # callbacks.run('on_val_batch_end')
+
+    # Compute metrics
+    stats = [torch.cat(x, 0).cpu().numpy() for x in zip(*stats)]  # to numpy
+    if len(stats) and stats[0].any():
+        results = ap_per_class_box_and_mask(*stats, plot=plots, save_dir=save_dir, names=names)
+        metrics.update(results)
+    nt = np.bincount(stats[4].astype(int), minlength=nc)  # number of targets per class
+
+    # Print results
+    pf = '%22s' + '%11i' * 2 + '%11.3g' * 8  # print format
+    LOGGER.info(pf % ("all", seen, nt.sum(), *metrics.mean_results()))
+    if nt.sum() == 0:
+        LOGGER.warning(f'WARNING ⚠️ no labels found in {task} set, can not compute metrics without labels')
+
+    # Print results per class
+    if (verbose or (nc < 50 and not training)) and nc > 1 and len(stats):
+        for i, c in enumerate(metrics.ap_class_index):
+            LOGGER.info(pf % (names[c], seen, nt[c], *metrics.class_result(i)))
+
+    # Print speeds
+    t = tuple(x.t / seen * 1E3 for x in dt)  # speeds per image
+    if not training:
+        shape = (batch_size, 3, imgsz, imgsz)
+        LOGGER.info(f'Speed: %.1fms pre-process, %.1fms inference, %.1fms NMS per image at shape {shape}' % t)
+
+    # Plots
+    if plots:
+        confusion_matrix.plot(save_dir=save_dir, names=list(names.values()))
+    # callbacks.run('on_val_end')
+
+    mp_bbox, mr_bbox, map50_bbox, map_bbox, mp_mask, mr_mask, map50_mask, map_mask = metrics.mean_results()
+
+    # Save JSON
+    if save_json and len(jdict):
+        w = Path(weights[0] if isinstance(weights, list) else weights).stem if weights is not None else ''  # weights
+        anno_json = str(Path(data.get('path', '../coco')) / 'annotations/instances_val2017.json')  # annotations json
+        pred_json = str(save_dir / f"{w}_predictions.json")  # predictions json
+        LOGGER.info(f'\nEvaluating pycocotools mAP... saving {pred_json}...')
+        with open(pred_json, 'w') as f:
+            json.dump(jdict, f)
+
+        try:  # https://github.com/cocodataset/cocoapi/blob/master/PythonAPI/pycocoEvalDemo.ipynb
+            from pycocotools.coco import COCO
+            from pycocotools.cocoeval import COCOeval
+
+            anno = COCO(anno_json)  # init annotations api
+            pred = anno.loadRes(pred_json)  # init predictions api
+            results = []
+            for eval in COCOeval(anno, pred, 'bbox'), COCOeval(anno, pred, 'segm'):
+                if is_coco:
+                    eval.params.imgIds = [int(Path(x).stem) for x in dataloader.dataset.im_files]  # img ID to evaluate
+                eval.evaluate()
+                eval.accumulate()
+                eval.summarize()
+                results.extend(eval.stats[:2])  # update results (mAP@0.5:0.95, mAP@0.5)
+            map_bbox, map50_bbox, map_mask, map50_mask = results
+        except Exception as e:
+            LOGGER.info(f'pycocotools unable to run: {e}')
+
+    # Return results
+    model.float()  # for training
+    if not training:
+        s = f"\n{len(list(save_dir.glob('labels/*.txt')))} labels saved to {save_dir / 'labels'}" if save_txt else ''
+        LOGGER.info(f"Results saved to {colorstr('bold', save_dir)}{s}")
+    final_metric = mp_bbox, mr_bbox, map50_bbox, map_bbox, mp_mask, mr_mask, map50_mask, map_mask
+    return (*final_metric, *(loss.cpu() / len(dataloader)).tolist()), metrics.get_maps(nc), t
+
+
+def parse_opt():
+    parser = argparse.ArgumentParser()
+    parser.add_argument('--data', type=str, default=ROOT / 'data/coco128-seg.yaml', help='dataset.yaml path')
+    parser.add_argument('--weights', nargs='+', type=str, default=ROOT / 'yolov5s-seg.pt', help='model path(s)')
+    parser.add_argument('--batch-size', type=int, default=32, help='batch size')
+    parser.add_argument('--imgsz', '--img', '--img-size', type=int, default=640, help='inference size (pixels)')
+    parser.add_argument('--conf-thres', type=float, default=0.001, help='confidence threshold')
+    parser.add_argument('--iou-thres', type=float, default=0.6, help='NMS IoU threshold')
+    parser.add_argument('--max-det', type=int, default=300, help='maximum detections per image')
+    parser.add_argument('--task', default='val', help='train, val, test, speed or study')
+    parser.add_argument('--device', default='', help='cuda device, i.e. 0 or 0,1,2,3 or cpu')
+    parser.add_argument('--workers', type=int, default=8, help='max dataloader workers (per RANK in DDP mode)')
+    parser.add_argument('--single-cls', action='store_true', help='treat as single-class dataset')
+    parser.add_argument('--augment', action='store_true', help='augmented inference')
+    parser.add_argument('--verbose', action='store_true', help='report mAP by class')
+    parser.add_argument('--save-txt', action='store_true', help='save results to *.txt')
+    parser.add_argument('--save-hybrid', action='store_true', help='save label+prediction hybrid results to *.txt')
+    parser.add_argument('--save-conf', action='store_true', help='save confidences in --save-txt labels')
+    parser.add_argument('--save-json', action='store_true', help='save a COCO-JSON results file')
+    parser.add_argument('--project', default=ROOT / 'runs/val-seg', help='save results to project/name')
+    parser.add_argument('--name', default='exp', help='save to project/name')
+    parser.add_argument('--exist-ok', action='store_true', help='existing project/name ok, do not increment')
+    parser.add_argument('--half', action='store_true', help='use FP16 half-precision inference')
+    parser.add_argument('--dnn', action='store_true', help='use OpenCV DNN for ONNX inference')
+    opt = parser.parse_args()
+    opt.data = check_yaml(opt.data)  # check YAML
+    # opt.save_json |= opt.data.endswith('coco.yaml')
+    opt.save_txt |= opt.save_hybrid
+    print_args(vars(opt))
+    return opt
+
+
+def main(opt):
+    check_requirements(requirements=ROOT / 'requirements.txt', exclude=('tensorboard', 'thop'))
+
+    if opt.task in ('train', 'val', 'test'):  # run normally
+        if opt.conf_thres > 0.001:  # https://github.com/ultralytics/yolov5/issues/1466
+            LOGGER.warning(f'WARNING ⚠️ confidence threshold {opt.conf_thres} > 0.001 produces invalid results')
+        if opt.save_hybrid:
+            LOGGER.warning('WARNING ⚠️ --save-hybrid returns high mAP from hybrid labels, not from predictions alone')
+        run(**vars(opt))
+
+    else:
+        weights = opt.weights if isinstance(opt.weights, list) else [opt.weights]
+        opt.half = torch.cuda.is_available() and opt.device != 'cpu'  # FP16 for fastest results
+        if opt.task == 'speed':  # speed benchmarks
+            # python val.py --task speed --data coco.yaml --batch 1 --weights yolov5n.pt yolov5s.pt...
+            opt.conf_thres, opt.iou_thres, opt.save_json = 0.25, 0.45, False
+            for opt.weights in weights:
+                run(**vars(opt), plots=False)
+
+        elif opt.task == 'study':  # speed vs mAP benchmarks
+            # python val.py --task study --data coco.yaml --iou 0.7 --weights yolov5n.pt yolov5s.pt...
+            for opt.weights in weights:
+                f = f'study_{Path(opt.data).stem}_{Path(opt.weights).stem}.txt'  # filename to save to
+                x, y = list(range(256, 1536 + 128, 128)), []  # x axis (image sizes), y axis
+                for opt.imgsz in x:  # img-size
+                    LOGGER.info(f'\nRunning {f} --imgsz {opt.imgsz}...')
+                    r, _, t = run(**vars(opt), plots=False)
+                    y.append(r + t)  # results and times
+                np.savetxt(f, y, fmt='%10.4g')  # save
+            os.system('zip -r study.zip study_*.txt')
+            plot_val_study(x=x)  # plot
+
+
+if __name__ == "__main__":
+    opt = parse_opt()
+    main(opt)

setup.cfg

+# Project-wide configuration file, can be used for package metadata and other toll configurations
+# Example usage: global configuration for PEP8 (via flake8) setting or default pytest arguments
+# Local usage: pip install pre-commit, pre-commit run --all-files
+
+[metadata]
+license_file = LICENSE
+description_file = README.md
+
+
+[tool:pytest]
+norecursedirs =
+    .git
+    dist
+    build
+addopts =
+    --doctest-modules
+    --durations=25
+    --color=yes
+
+
+[flake8]
+max-line-length = 120
+exclude = .tox,*.egg,build,temp
+select = E,W,F
+doctests = True
+verbose = 2
+# https://pep8.readthedocs.io/en/latest/intro.html#error-codes
+format = pylint
+# see: https://www.flake8rules.com/
+ignore =
+    E731  # Do not assign a lambda expression, use a def
+    F405  # name may be undefined, or defined from star imports: module
+    E402  # module level import not at top of file
+    F401  # module imported but unused
+    W504  # line break after binary operator
+    E127  # continuation line over-indented for visual indent
+    E231  # missing whitespace after ‘,’, ‘;’, or ‘:’
+    E501  # line too long
+    F403  # ‘from module import *’ used; unable to detect undefined names
+
+
+[isort]
+# https://pycqa.github.io/isort/docs/configuration/options.html
+line_length = 120
+# see: https://pycqa.github.io/isort/docs/configuration/multi_line_output_modes.html
+multi_line_output = 0
+
+
+[yapf]
+based_on_style = pep8
+spaces_before_comment = 2
+COLUMN_LIMIT = 120
+COALESCE_BRACKETS = True
+SPACES_AROUND_POWER_OPERATOR = True
+SPACE_BETWEEN_ENDING_COMMA_AND_CLOSING_BRACKET = False
+SPLIT_BEFORE_CLOSING_BRACKET = False
+SPLIT_BEFORE_FIRST_ARGUMENT = False
+# EACH_DICT_ENTRY_ON_SEPARATE_LINE = False

train.py

+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+"""
+Train a YOLOv5 model on a custom dataset.
+Models and datasets download automatically from the latest YOLOv5 release.
+
+Usage - Single-GPU training:
+    $ python train.py --data coco128.yaml --weights yolov5s.pt --img 640  # from pretrained (recommended)
+    $ python train.py --data coco128.yaml --weights '' --cfg yolov5s.yaml --img 640  # from scratch
+
+Usage - Multi-GPU DDP training:
+    $ python -m torch.distributed.run --nproc_per_node 4 --master_port 1 train.py --data coco128.yaml --weights yolov5s.pt --img 640 --device 0,1,2,3
+
+Models:     https://github.com/ultralytics/yolov5/tree/master/models
+Datasets:   https://github.com/ultralytics/yolov5/tree/master/data
+Tutorial:   https://github.com/ultralytics/yolov5/wiki/Train-Custom-Data
+"""
+
+import argparse
+import math
+import os
+import random
+import sys
+import time
+from copy import deepcopy
+from datetime import datetime
+from pathlib import Path
+
+import numpy as np
+import torch
+import torch.distributed as dist
+import torch.nn as nn
+import yaml
+from torch.optim import lr_scheduler
+from tqdm import tqdm
+
+FILE = Path(__file__).resolve()
+ROOT = FILE.parents[0]  # YOLOv5 root directory
+if str(ROOT) not in sys.path:
+    sys.path.append(str(ROOT))  # add ROOT to PATH
+ROOT = Path(os.path.relpath(ROOT, Path.cwd()))  # relative
+
+import val as validate  # for end-of-epoch mAP
+from models.experimental import attempt_load
+from models.yolo import Model
+from utils.autoanchor import check_anchors
+from utils.autobatch import check_train_batch_size
+from utils.callbacks import Callbacks
+from utils.dataloaders import create_dataloader
+from utils.downloads import attempt_download, is_url
+from utils.general import (LOGGER, TQDM_BAR_FORMAT, check_amp, check_dataset, check_file, check_git_info,
+                           check_git_status, check_img_size, check_requirements, check_suffix, check_yaml, colorstr,
+                           get_latest_run, increment_path, init_seeds, intersect_dicts, labels_to_class_weights,
+                           labels_to_image_weights, methods, one_cycle, print_args, print_mutation, strip_optimizer,
+                           yaml_save)
+from utils.loggers import Loggers
+from utils.loggers.comet.comet_utils import check_comet_resume
+from utils.loss import ComputeLoss
+from utils.metrics import fitness
+from utils.plots import plot_evolve
+from utils.torch_utils import (EarlyStopping, ModelEMA, de_parallel, select_device, smart_DDP, smart_optimizer,
+                               smart_resume, torch_distributed_zero_first)
+
+LOCAL_RANK = int(os.getenv('LOCAL_RANK', -1))  # https://pytorch.org/docs/stable/elastic/run.html
+RANK = int(os.getenv('RANK', -1))
+WORLD_SIZE = int(os.getenv('WORLD_SIZE', 1))
+GIT_INFO = check_git_info()
+
+
+def train(hyp, opt, device, callbacks):  # hyp is path/to/hyp.yaml or hyp dictionary
+    save_dir, epochs, batch_size, weights, single_cls, evolve, data, cfg, resume, noval, nosave, workers, freeze = \
+        Path(opt.save_dir), opt.epochs, opt.batch_size, opt.weights, opt.single_cls, opt.evolve, opt.data, opt.cfg, \
+        opt.resume, opt.noval, opt.nosave, opt.workers, opt.freeze
+    callbacks.run('on_pretrain_routine_start')
+
+    # Directories
+    w = save_dir / 'weights'  # weights dir
+    (w.parent if evolve else w).mkdir(parents=True, exist_ok=True)  # make dir
+    last, best = w / 'last.pt', w / 'best.pt'
+
+    # Hyperparameters
+    if isinstance(hyp, str):
+        with open(hyp, errors='ignore') as f:
+            hyp = yaml.safe_load(f)  # load hyps dict
+    LOGGER.info(colorstr('hyperparameters: ') + ', '.join(f'{k}={v}' for k, v in hyp.items()))
+    opt.hyp = hyp.copy()  # for saving hyps to checkpoints
+
+    # Save run settings
+    if not evolve:
+        yaml_save(save_dir / 'hyp.yaml', hyp)
+        yaml_save(save_dir / 'opt.yaml', vars(opt))
+
+    # Loggers
+    data_dict = None
+    if RANK in {-1, 0}:
+        loggers = Loggers(save_dir, weights, opt, hyp, LOGGER)  # loggers instance
+
+        # Register actions
+        for k in methods(loggers):
+            callbacks.register_action(k, callback=getattr(loggers, k))
+
+        # Process custom dataset artifact link
+        data_dict = loggers.remote_dataset
+        if resume:  # If resuming runs from remote artifact
+            weights, epochs, hyp, batch_size = opt.weights, opt.epochs, opt.hyp, opt.batch_size
+
+    # Config
+    plots = not evolve and not opt.noplots  # create plots
+    cuda = device.type != 'cpu'
+    init_seeds(opt.seed + 1 + RANK, deterministic=False)
+    with torch_distributed_zero_first(LOCAL_RANK):
+        data_dict = data_dict or check_dataset(data)  # check if None
+    train_path, val_path = data_dict['train'], data_dict['val']
+    nc = 1 if single_cls else int(data_dict['nc'])  # number of classes
+    names = {0: 'item'} if single_cls and len(data_dict['names']) != 1 else data_dict['names']  # class names
+    is_coco = isinstance(val_path, str) and val_path.endswith('coco/val2017.txt')  # COCO dataset
+
+    # Model
+    check_suffix(weights, '.pt')  # check weights
+    pretrained = weights.endswith('.pt')
+    if pretrained:
+        with torch_distributed_zero_first(LOCAL_RANK):
+            weights = attempt_download(weights)  # download if not found locally
+        ckpt = torch.load(weights, map_location='cpu')  # load checkpoint to CPU to avoid CUDA memory leak
+        model = Model(cfg or ckpt['model'].yaml, ch=3, nc=nc, anchors=hyp.get('anchors')).to(device)  # create
+        exclude = ['anchor'] if (cfg or hyp.get('anchors')) and not resume else []  # exclude keys
+        csd = ckpt['model'].float().state_dict()  # checkpoint state_dict as FP32
+        csd = intersect_dicts(csd, model.state_dict(), exclude=exclude)  # intersect
+        model.load_state_dict(csd, strict=False)  # load
+        LOGGER.info(f'Transferred {len(csd)}/{len(model.state_dict())} items from {weights}')  # report
+    else:
+        model = Model(cfg, ch=3, nc=nc, anchors=hyp.get('anchors')).to(device)  # create
+    amp = check_amp(model)  # check AMP
+
+    # Freeze
+    freeze = [f'model.{x}.' for x in (freeze if len(freeze) > 1 else range(freeze[0]))]  # layers to freeze
+    for k, v in model.named_parameters():
+        v.requires_grad = True  # train all layers
+        # v.register_hook(lambda x: torch.nan_to_num(x))  # NaN to 0 (commented for erratic training results)
+        if any(x in k for x in freeze):
+            LOGGER.info(f'freezing {k}')
+            v.requires_grad = False
+
+    # Image size
+    gs = max(int(model.stride.max()), 32)  # grid size (max stride)
+    imgsz = check_img_size(opt.imgsz, gs, floor=gs * 2)  # verify imgsz is gs-multiple
+
+    # Batch size
+    if RANK == -1 and batch_size == -1:  # single-GPU only, estimate best batch size
+        batch_size = check_train_batch_size(model, imgsz, amp)
+        loggers.on_params_update({"batch_size": batch_size})
+
+    # Optimizer
+    nbs = 64  # nominal batch size
+    accumulate = max(round(nbs / batch_size), 1)  # accumulate loss before optimizing
+    hyp['weight_decay'] *= batch_size * accumulate / nbs  # scale weight_decay
+    optimizer = smart_optimizer(model, opt.optimizer, hyp['lr0'], hyp['momentum'], hyp['weight_decay'])
+
+    # Scheduler
+    if opt.cos_lr:
+        lf = one_cycle(1, hyp['lrf'], epochs)  # cosine 1->hyp['lrf']
+    else:
+        lf = lambda x: (1 - x / epochs) * (1.0 - hyp['lrf']) + hyp['lrf']  # linear
+    scheduler = lr_scheduler.LambdaLR(optimizer, lr_lambda=lf)  # plot_lr_scheduler(optimizer, scheduler, epochs)
+
+    # EMA
+    ema = ModelEMA(model) if RANK in {-1, 0} else None
+
+    # Resume
+    best_fitness, start_epoch = 0.0, 0
+    if pretrained:
+        if resume:
+            best_fitness, start_epoch, epochs = smart_resume(ckpt, optimizer, ema, weights, epochs, resume)
+        del ckpt, csd
+
+    # DP mode
+    if cuda and RANK == -1 and torch.cuda.device_count() > 1:
+        LOGGER.warning('WARNING ⚠️ DP not recommended, use torch.distributed.run for best DDP Multi-GPU results.\n'
+                       'See Multi-GPU Tutorial at https://github.com/ultralytics/yolov5/issues/475 to get started.')
+        model = torch.nn.DataParallel(model)
+
+    # SyncBatchNorm
+    if opt.sync_bn and cuda and RANK != -1:
+        model = torch.nn.SyncBatchNorm.convert_sync_batchnorm(model).to(device)
+        LOGGER.info('Using SyncBatchNorm()')
+
+    # Trainloader
+    train_loader, dataset = create_dataloader(train_path,
+                                              imgsz,
+                                              batch_size // WORLD_SIZE,
+                                              gs,
+                                              single_cls,
+                                              hyp=hyp,
+                                              augment=True,
+                                              cache=None if opt.cache == 'val' else opt.cache,
+                                              rect=opt.rect,
+                                              rank=LOCAL_RANK,
+                                              workers=workers,
+                                              image_weights=opt.image_weights,
+                                              quad=opt.quad,
+                                              prefix=colorstr('train: '),
+                                              shuffle=True)
+    labels = np.concatenate(dataset.labels, 0)
+    mlc = int(labels[:, 0].max())  # max label class
+    assert mlc < nc, f'Label class {mlc} exceeds nc={nc} in {data}. Possible class labels are 0-{nc - 1}'
+
+    # Process 0
+    if RANK in {-1, 0}:
+        val_loader = create_dataloader(val_path,
+                                       imgsz,
+                                       batch_size // WORLD_SIZE * 2,
+                                       gs,
+                                       single_cls,
+                                       hyp=hyp,
+                                       cache=None if noval else opt.cache,
+                                       rect=True,
+                                       rank=-1,
+                                       workers=workers * 2,
+                                       pad=0.5,
+                                       prefix=colorstr('val: '))[0]
+
+        if not resume:
+            if not opt.noautoanchor:
+                check_anchors(dataset, model=model, thr=hyp['anchor_t'], imgsz=imgsz)  # run AutoAnchor
+            model.half().float()  # pre-reduce anchor precision
+
+        callbacks.run('on_pretrain_routine_end', labels, names)
+
+    # DDP mode
+    if cuda and RANK != -1:
+        model = smart_DDP(model)
+
+    # Model attributes
+    nl = de_parallel(model).model[-1].nl  # number of detection layers (to scale hyps)
+    hyp['box'] *= 3 / nl  # scale to layers
+    hyp['cls'] *= nc / 80 * 3 / nl  # scale to classes and layers
+    hyp['obj'] *= (imgsz / 640) ** 2 * 3 / nl  # scale to image size and layers
+    hyp['label_smoothing'] = opt.label_smoothing
+    model.nc = nc  # attach number of classes to model
+    model.hyp = hyp  # attach hyperparameters to model
+    model.class_weights = labels_to_class_weights(dataset.labels, nc).to(device) * nc  # attach class weights
+    model.names = names
+
+    # Start training
+    t0 = time.time()
+    nb = len(train_loader)  # number of batches
+    nw = max(round(hyp['warmup_epochs'] * nb), 100)  # number of warmup iterations, max(3 epochs, 100 iterations)
+    # nw = min(nw, (epochs - start_epoch) / 2 * nb)  # limit warmup to < 1/2 of training
+    last_opt_step = -1
+    maps = np.zeros(nc)  # mAP per class
+    results = (0, 0, 0, 0, 0, 0, 0)  # P, R, mAP@.5, mAP@.5-.95, val_loss(box, obj, cls)
+    scheduler.last_epoch = start_epoch - 1  # do not move
+    scaler = torch.cuda.amp.GradScaler(enabled=amp)
+    stopper, stop = EarlyStopping(patience=opt.patience), False
+    compute_loss = ComputeLoss(model)  # init loss class
+    callbacks.run('on_train_start')
+    LOGGER.info(f'Image sizes {imgsz} train, {imgsz} val\n'
+                f'Using {train_loader.num_workers * WORLD_SIZE} dataloader workers\n'
+                f"Logging results to {colorstr('bold', save_dir)}\n"
+                f'Starting training for {epochs} epochs...')
+    for epoch in range(start_epoch, epochs):  # epoch ------------------------------------------------------------------
+        callbacks.run('on_train_epoch_start')
+        model.train()
+
+        # Update image weights (optional, single-GPU only)
+        if opt.image_weights:
+            cw = model.class_weights.cpu().numpy() * (1 - maps) ** 2 / nc  # class weights
+            iw = labels_to_image_weights(dataset.labels, nc=nc, class_weights=cw)  # image weights
+            dataset.indices = random.choices(range(dataset.n), weights=iw, k=dataset.n)  # rand weighted idx
+
+        # Update mosaic border (optional)
+        # b = int(random.uniform(0.25 * imgsz, 0.75 * imgsz + gs) // gs * gs)
+        # dataset.mosaic_border = [b - imgsz, -b]  # height, width borders
+
+        mloss = torch.zeros(3, device=device)  # mean losses
+        if RANK != -1:
+            train_loader.sampler.set_epoch(epoch)
+        pbar = enumerate(train_loader)
+        LOGGER.info(('\n' + '%11s' * 7) % ('Epoch', 'GPU_mem', 'box_loss', 'obj_loss', 'cls_loss', 'Instances', 'Size'))
+        if RANK in {-1, 0}:
+            pbar = tqdm(pbar, total=nb, bar_format=TQDM_BAR_FORMAT)  # progress bar
+        optimizer.zero_grad()
+        for i, (imgs, targets, paths, _) in pbar:  # batch -------------------------------------------------------------
+            callbacks.run('on_train_batch_start')
+            ni = i + nb * epoch  # number integrated batches (since train start)
+            imgs = imgs.to(device, non_blocking=True).float() / 255  # uint8 to float32, 0-255 to 0.0-1.0
+
+            # Warmup
+            if ni <= nw:
+                xi = [0, nw]  # x interp
+                # compute_loss.gr = np.interp(ni, xi, [0.0, 1.0])  # iou loss ratio (obj_loss = 1.0 or iou)
+                accumulate = max(1, np.interp(ni, xi, [1, nbs / batch_size]).round())
+                for j, x in enumerate(optimizer.param_groups):
+                    # bias lr falls from 0.1 to lr0, all other lrs rise from 0.0 to lr0
+                    x['lr'] = np.interp(ni, xi, [hyp['warmup_bias_lr'] if j == 0 else 0.0, x['initial_lr'] * lf(epoch)])
+                    if 'momentum' in x:
+                        x['momentum'] = np.interp(ni, xi, [hyp['warmup_momentum'], hyp['momentum']])
+
+            # Multi-scale
+            if opt.multi_scale:
+                sz = random.randrange(imgsz * 0.5, imgsz * 1.5 + gs) // gs * gs  # size
+                sf = sz / max(imgs.shape[2:])  # scale factor
+                if sf != 1:
+                    ns = [math.ceil(x * sf / gs) * gs for x in imgs.shape[2:]]  # new shape (stretched to gs-multiple)
+                    imgs = nn.functional.interpolate(imgs, size=ns, mode='bilinear', align_corners=False)
+
+            # Forward
+            with torch.amp.autocast(device_type='cuda', enabled=amp):
+                pred = model(imgs)  # forward
+                loss, loss_items = compute_loss(pred, targets.to(device))  # loss scaled by batch_size
+                if RANK != -1:
+                    loss *= WORLD_SIZE  # gradient averaged between devices in DDP mode
+                if opt.quad:
+                    loss *= 4.
+
+            # Backward
+            scaler.scale(loss).backward()
+
+            # Optimize - https://pytorch.org/docs/master/notes/amp_examples.html
+            if ni - last_opt_step >= accumulate:
+                scaler.unscale_(optimizer)  # unscale gradients
+                torch.nn.utils.clip_grad_norm_(model.parameters(), max_norm=10.0)  # clip gradients
+                scaler.step(optimizer)  # optimizer.step
+                scaler.update()
+                optimizer.zero_grad()
+                if ema:
+                    ema.update(model)
+                last_opt_step = ni
+
+            # Log
+            if RANK in {-1, 0}:
+                mloss = (mloss * i + loss_items) / (i + 1)  # update mean losses
+                mem = f'{torch.cuda.memory_reserved() / 1E9 if torch.cuda.is_available() else 0:.3g}G'  # (GB)
+                pbar.set_description(('%11s' * 2 + '%11.4g' * 5) %
+                                     (f'{epoch}/{epochs - 1}', mem, *mloss, targets.shape[0], imgs.shape[-1]))
+                callbacks.run('on_train_batch_end', model, ni, imgs, targets, paths, list(mloss))
+                if callbacks.stop_training:
+                    return
+            # end batch ------------------------------------------------------------------------------------------------
+
+        # Scheduler
+        lr = [x['lr'] for x in optimizer.param_groups]  # for loggers
+        scheduler.step()
+
+        if RANK in {-1, 0}:
+            # mAP
+            callbacks.run('on_train_epoch_end', epoch=epoch)
+            ema.update_attr(model, include=['yaml', 'nc', 'hyp', 'names', 'stride', 'class_weights'])
+            final_epoch = (epoch + 1 == epochs) or stopper.possible_stop
+            if not noval or final_epoch:  # Calculate mAP
+                results, maps, _ = validate.run(data_dict,
+                                                batch_size=batch_size // WORLD_SIZE * 2,
+                                                imgsz=imgsz,
+                                                half=amp,
+                                                model=ema.ema,
+                                                single_cls=single_cls,
+                                                dataloader=val_loader,
+                                                save_dir=save_dir,
+                                                plots=False,
+                                                callbacks=callbacks,
+                                                compute_loss=compute_loss)
+
+            # Update best mAP
+            fi = fitness(np.array(results).reshape(1, -1))  # weighted combination of [P, R, mAP@.5, mAP@.5-.95]
+            stop = stopper(epoch=epoch, fitness=fi)  # early stop check
+            if fi > best_fitness:
+                best_fitness = fi
+            log_vals = list(mloss) + list(results) + lr
+            callbacks.run('on_fit_epoch_end', log_vals, epoch, best_fitness, fi)
+
+            # Save model
+            if (not nosave) or (final_epoch and not evolve):  # if save
+                ckpt = {
+                    'epoch': epoch,
+                    'best_fitness': best_fitness,
+                    'model': deepcopy(de_parallel(model)).half(),
+                    'ema': deepcopy(ema.ema).half(),
+                    'updates': ema.updates,
+                    'optimizer': optimizer.state_dict(),
+                    'opt': vars(opt),
+                    'git': GIT_INFO,  # {remote, branch, commit} if a git repo
+                    'date': datetime.now().isoformat()}
+
+                # Save last, best and delete
+                torch.save(ckpt, last)
+                if best_fitness == fi:
+                    torch.save(ckpt, best)
+                if opt.save_period > 0 and epoch % opt.save_period == 0:
+                    torch.save(ckpt, w / f'epoch{epoch}.pt')
+                del ckpt
+                callbacks.run('on_model_save', last, epoch, final_epoch, best_fitness, fi)
+
+        # EarlyStopping
+        if RANK != -1:  # if DDP training
+            broadcast_list = [stop if RANK == 0 else None]
+            dist.broadcast_object_list(broadcast_list, 0)  # broadcast 'stop' to all ranks
+            if RANK != 0:
+                stop = broadcast_list[0]
+        if stop:
+            break  # must break all DDP ranks
+
+        # end epoch ----------------------------------------------------------------------------------------------------
+    # end training -----------------------------------------------------------------------------------------------------
+    if RANK in {-1, 0}:
+        LOGGER.info(f'\n{epoch - start_epoch + 1} epochs completed in {(time.time() - t0) / 3600:.3f} hours.')
+        for f in last, best:
+            if f.exists():
+                strip_optimizer(f)  # strip optimizers
+                if f is best:
+                    LOGGER.info(f'\nValidating {f}...')
+                    results, _, _ = validate.run(
+                        data_dict,
+                        batch_size=batch_size // WORLD_SIZE * 2,
+                        imgsz=imgsz,
+                        model=attempt_load(f, device).half(),
+                        iou_thres=0.65 if is_coco else 0.60,  # best pycocotools at iou 0.65
+                        single_cls=single_cls,
+                        dataloader=val_loader,
+                        save_dir=save_dir,
+                        save_json=is_coco,
+                        verbose=True,
+                        plots=plots,
+                        callbacks=callbacks,
+                        compute_loss=compute_loss)  # val best model with plots
+                    if is_coco:
+                        callbacks.run('on_fit_epoch_end', list(mloss) + list(results) + lr, epoch, best_fitness, fi)
+
+        callbacks.run('on_train_end', last, best, epoch, results)
+
+    torch.cuda.empty_cache()
+    return results
+
+
+def parse_opt(known=False):
+    parser = argparse.ArgumentParser()
+    parser.add_argument('--weights', type=str, default=ROOT / 'yolov5s.pt', help='initial weights path')
+    parser.add_argument('--cfg', type=str, default='', help='model.yaml path')
+    parser.add_argument('--data', type=str, default=ROOT / 'data/coco128.yaml', help='dataset.yaml path')
+    parser.add_argument('--hyp', type=str, default=ROOT / 'data/hyps/hyp.scratch-low.yaml', help='hyperparameters path')
+    parser.add_argument('--epochs', type=int, default=100, help='total training epochs')
+    parser.add_argument('--batch-size', type=int, default=16, help='total batch size for all GPUs, -1 for autobatch')
+    parser.add_argument('--imgsz', '--img', '--img-size', type=int, default=640, help='train, val image size (pixels)')
+    parser.add_argument('--rect', action='store_true', help='rectangular training')
+    parser.add_argument('--resume', nargs='?', const=True, default=False, help='resume most recent training')
+    parser.add_argument('--nosave', action='store_true', help='only save final checkpoint')
+    parser.add_argument('--noval', action='store_true', help='only validate final epoch')
+    parser.add_argument('--noautoanchor', action='store_true', help='disable AutoAnchor')
+    parser.add_argument('--noplots', action='store_true', help='save no plot files')
+    parser.add_argument('--evolve', type=int, nargs='?', const=300, help='evolve hyperparameters for x generations')
+    parser.add_argument('--bucket', type=str, default='', help='gsutil bucket')
+    parser.add_argument('--cache', type=str, nargs='?', const='ram', help='image --cache ram/disk')
+    parser.add_argument('--image-weights', action='store_true', help='use weighted image selection for training')
+    parser.add_argument('--device', default='', help='cuda device, i.e. 0 or 0,1,2,3 or cpu')
+    parser.add_argument('--multi-scale', action='store_true', help='vary img-size +/- 50%%')
+    parser.add_argument('--single-cls', action='store_true', help='train multi-class data as single-class')
+    parser.add_argument('--optimizer', type=str, choices=['SGD', 'Adam', 'AdamW'], default='SGD', help='optimizer')
+    parser.add_argument('--sync-bn', action='store_true', help='use SyncBatchNorm, only available in DDP mode')
+    parser.add_argument('--workers', type=int, default=8, help='max dataloader workers (per RANK in DDP mode)')
+    parser.add_argument('--project', default=ROOT / 'runs/train', help='save to project/name')
+    parser.add_argument('--name', default='exp', help='save to project/name')
+    parser.add_argument('--exist-ok', action='store_true', help='existing project/name ok, do not increment')
+    parser.add_argument('--quad', action='store_true', help='quad dataloader')
+    parser.add_argument('--cos-lr', action='store_true', help='cosine LR scheduler')
+    parser.add_argument('--label-smoothing', type=float, default=0.0, help='Label smoothing epsilon')
+    parser.add_argument('--patience', type=int, default=100, help='EarlyStopping patience (epochs without improvement)')
+    parser.add_argument('--freeze', nargs='+', type=int, default=[0], help='Freeze layers: backbone=10, first3=0 1 2')
+    parser.add_argument('--save-period', type=int, default=-1, help='Save checkpoint every x epochs (disabled if < 1)')
+    parser.add_argument('--seed', type=int, default=0, help='Global training seed')
+    parser.add_argument('--local_rank', type=int, default=-1, help='Automatic DDP Multi-GPU argument, do not modify')
+
+    # Logger arguments
+    parser.add_argument('--entity', default=None, help='Entity')
+    parser.add_argument('--upload_dataset', nargs='?', const=True, default=False, help='Upload data, "val" option')
+    parser.add_argument('--bbox_interval', type=int, default=-1, help='Set bounding-box image logging interval')
+    parser.add_argument('--artifact_alias', type=str, default='latest', help='Version of dataset artifact to use')
+
+    return parser.parse_known_args()[0] if known else parser.parse_args()
+
+
+def main(opt, callbacks=Callbacks()):
+    # Checks
+    if RANK in {-1, 0}:
+        print_args(vars(opt))
+        check_git_status()
+        check_requirements()
+
+    # Resume (from specified or most recent last.pt)
+    if opt.resume and not check_comet_resume(opt) and not opt.evolve:
+        last = Path(check_file(opt.resume) if isinstance(opt.resume, str) else get_latest_run())
+        opt_yaml = last.parent.parent / 'opt.yaml'  # train options yaml
+        opt_data = opt.data  # original dataset
+        if opt_yaml.is_file():
+            with open(opt_yaml, errors='ignore') as f:
+                d = yaml.safe_load(f)
+        else:
+            d = torch.load(last, map_location='cpu')['opt']
+        opt = argparse.Namespace(**d)  # replace
+        opt.cfg, opt.weights, opt.resume = '', str(last), True  # reinstate
+        if is_url(opt_data):
+            opt.data = check_file(opt_data)  # avoid HUB resume auth timeout
+    else:
+        opt.data, opt.cfg, opt.hyp, opt.weights, opt.project = \
+            check_file(opt.data), check_yaml(opt.cfg), check_yaml(opt.hyp), str(opt.weights), str(opt.project)  # checks
+        assert len(opt.cfg) or len(opt.weights), 'either --cfg or --weights must be specified'
+        if opt.evolve:
+            if opt.project == str(ROOT / 'runs/train'):  # if default project name, rename to runs/evolve
+                opt.project = str(ROOT / 'runs/evolve')
+            opt.exist_ok, opt.resume = opt.resume, False  # pass resume to exist_ok and disable resume
+        if opt.name == 'cfg':
+            opt.name = Path(opt.cfg).stem  # use model.yaml as name
+        opt.save_dir = str(increment_path(Path(opt.project) / opt.name, exist_ok=opt.exist_ok))
+
+    # DDP mode
+    device = select_device(opt.device, batch_size=opt.batch_size)
+    if LOCAL_RANK != -1:
+        msg = 'is not compatible with YOLOv5 Multi-GPU DDP training'
+        assert not opt.image_weights, f'--image-weights {msg}'
+        assert not opt.evolve, f'--evolve {msg}'
+        assert opt.batch_size != -1, f'AutoBatch with --batch-size -1 {msg}, please pass a valid --batch-size'
+        assert opt.batch_size % WORLD_SIZE == 0, f'--batch-size {opt.batch_size} must be multiple of WORLD_SIZE'
+        assert torch.cuda.device_count() > LOCAL_RANK, 'insufficient CUDA devices for DDP command'
+        torch.cuda.set_device(LOCAL_RANK)
+        device = torch.device('cuda', LOCAL_RANK)
+        dist.init_process_group(backend="nccl" if dist.is_nccl_available() else "gloo")
+
+    # Train
+    if not opt.evolve:
+        train(opt.hyp, opt, device, callbacks)
+
+    # Evolve hyperparameters (optional)
+    else:
+        # Hyperparameter evolution metadata (mutation scale 0-1, lower_limit, upper_limit)
+        meta = {
+            'lr0': (1, 1e-5, 1e-1),  # initial learning rate (SGD=1E-2, Adam=1E-3)
+            'lrf': (1, 0.01, 1.0),  # final OneCycleLR learning rate (lr0 * lrf)
+            'momentum': (0.3, 0.6, 0.98),  # SGD momentum/Adam beta1
+            'weight_decay': (1, 0.0, 0.001),  # optimizer weight decay
+            'warmup_epochs': (1, 0.0, 5.0),  # warmup epochs (fractions ok)
+            'warmup_momentum': (1, 0.0, 0.95),  # warmup initial momentum
+            'warmup_bias_lr': (1, 0.0, 0.2),  # warmup initial bias lr
+            'box': (1, 0.02, 0.2),  # box loss gain
+            'cls': (1, 0.2, 4.0),  # cls loss gain
+            'cls_pw': (1, 0.5, 2.0),  # cls BCELoss positive_weight
+            'obj': (1, 0.2, 4.0),  # obj loss gain (scale with pixels)
+            'obj_pw': (1, 0.5, 2.0),  # obj BCELoss positive_weight
+            'iou_t': (0, 0.1, 0.7),  # IoU training threshold
+            'anchor_t': (1, 2.0, 8.0),  # anchor-multiple threshold
+            'anchors': (2, 2.0, 10.0),  # anchors per output grid (0 to ignore)
+            'fl_gamma': (0, 0.0, 2.0),  # focal loss gamma (efficientDet default gamma=1.5)
+            'hsv_h': (1, 0.0, 0.1),  # image HSV-Hue augmentation (fraction)
+            'hsv_s': (1, 0.0, 0.9),  # image HSV-Saturation augmentation (fraction)
+            'hsv_v': (1, 0.0, 0.9),  # image HSV-Value augmentation (fraction)
+            'degrees': (1, 0.0, 45.0),  # image rotation (+/- deg)
+            'translate': (1, 0.0, 0.9),  # image translation (+/- fraction)
+            'scale': (1, 0.0, 0.9),  # image scale (+/- gain)
+            'shear': (1, 0.0, 10.0),  # image shear (+/- deg)
+            'perspective': (0, 0.0, 0.001),  # image perspective (+/- fraction), range 0-0.001
+            'flipud': (1, 0.0, 1.0),  # image flip up-down (probability)
+            'fliplr': (0, 0.0, 1.0),  # image flip left-right (probability)
+            'mosaic': (1, 0.0, 1.0),  # image mixup (probability)
+            'mixup': (1, 0.0, 1.0),  # image mixup (probability)
+            'copy_paste': (1, 0.0, 1.0)}  # segment copy-paste (probability)
+
+        with open(opt.hyp, errors='ignore') as f:
+            hyp = yaml.safe_load(f)  # load hyps dict
+            if 'anchors' not in hyp:  # anchors commented in hyp.yaml
+                hyp['anchors'] = 3
+        if opt.noautoanchor:
+            del hyp['anchors'], meta['anchors']
+        opt.noval, opt.nosave, save_dir = True, True, Path(opt.save_dir)  # only val/save final epoch
+        # ei = [isinstance(x, (int, float)) for x in hyp.values()]  # evolvable indices
+        evolve_yaml, evolve_csv = save_dir / 'hyp_evolve.yaml', save_dir / 'evolve.csv'
+        if opt.bucket:
+            os.system(f'gsutil cp gs://{opt.bucket}/evolve.csv {evolve_csv}')  # download evolve.csv if exists
+
+        for _ in range(opt.evolve):  # generations to evolve
+            if evolve_csv.exists():  # if evolve.csv exists: select best hyps and mutate
+                # Select parent(s)
+                parent = 'single'  # parent selection method: 'single' or 'weighted'
+                x = np.loadtxt(evolve_csv, ndmin=2, delimiter=',', skiprows=1)
+                n = min(5, len(x))  # number of previous results to consider
+                x = x[np.argsort(-fitness(x))][:n]  # top n mutations
+                w = fitness(x) - fitness(x).min() + 1E-6  # weights (sum > 0)
+                if parent == 'single' or len(x) == 1:
+                    # x = x[random.randint(0, n - 1)]  # random selection
+                    x = x[random.choices(range(n), weights=w)[0]]  # weighted selection
+                elif parent == 'weighted':
+                    x = (x * w.reshape(n, 1)).sum(0) / w.sum()  # weighted combination
+
+                # Mutate
+                mp, s = 0.8, 0.2  # mutation probability, sigma
+                npr = np.random
+                npr.seed(int(time.time()))
+                g = np.array([meta[k][0] for k in hyp.keys()])  # gains 0-1
+                ng = len(meta)
+                v = np.ones(ng)
+                while all(v == 1):  # mutate until a change occurs (prevent duplicates)
+                    v = (g * (npr.random(ng) < mp) * npr.randn(ng) * npr.random() * s + 1).clip(0.3, 3.0)
+                for i, k in enumerate(hyp.keys()):  # plt.hist(v.ravel(), 300)
+                    hyp[k] = float(x[i + 7] * v[i])  # mutate
+
+            # Constrain to limits
+            for k, v in meta.items():
+                hyp[k] = max(hyp[k], v[1])  # lower limit
+                hyp[k] = min(hyp[k], v[2])  # upper limit
+                hyp[k] = round(hyp[k], 5)  # significant digits
+
+            # Train mutation
+            results = train(hyp.copy(), opt, device, callbacks)
+            callbacks = Callbacks()
+            # Write mutation results
+            keys = ('metrics/precision', 'metrics/recall', 'metrics/mAP_0.5', 'metrics/mAP_0.5:0.95', 'val/box_loss',
+                    'val/obj_loss', 'val/cls_loss')
+            print_mutation(keys, results, hyp.copy(), save_dir, opt.bucket)
+
+        # Plot results
+        plot_evolve(evolve_csv)
+        LOGGER.info(f'Hyperparameter evolution finished {opt.evolve} generations\n'
+                    f"Results saved to {colorstr('bold', save_dir)}\n"
+                    f'Usage example: $ python train.py --hyp {evolve_yaml}')
+
+
+def run(**kwargs):
+    # Usage: import train; train.run(data='coco128.yaml', imgsz=320, weights='yolov5m.pt')
+    opt = parse_opt(True)
+    for k, v in kwargs.items():
+        setattr(opt, k, v)
+    main(opt)
+    return opt
+
+
+if __name__ == "__main__":
+    opt = parse_opt()
+    main(opt)

tutorial.ipynb

+{
+  "nbformat": 4,
+  "nbformat_minor": 0,
+  "metadata": {
+    "colab": {
+      "name": "YOLOv5 Tutorial",
+      "provenance": [],
+      "toc_visible": true
+    },
+    "kernelspec": {
+      "name": "python3",
+      "display_name": "Python 3"
+    },
+    "accelerator": "GPU",
+    "widgets": {
+      "application/vnd.jupyter.widget-state+json": {
+        "300b4d5355ef4967bd5246afeef6eef5": {
+          "model_module": "@jupyter-widgets/controls",
+          "model_name": "HBoxModel",
+          "model_module_version": "1.5.0",
+          "state": {
+            "_dom_classes": [],
+            "_model_module": "@jupyter-widgets/controls",
+            "_model_module_version": "1.5.0",
+            "_model_name": "HBoxModel",
+            "_view_count": null,
+            "_view_module": "@jupyter-widgets/controls",
+            "_view_module_version": "1.5.0",
+            "_view_name": "HBoxView",
+            "box_style": "",
+            "children": [
+              "IPY_MODEL_84e6829bb88845a8a4f42700b8496925",
+              "IPY_MODEL_c038e52d41bf4d5b9602930c3d074087",
+              "IPY_MODEL_2667604641764341b0bc8c6afea438fd"
+            ],
+            "layout": "IPY_MODEL_98b3a4806ed14102b0d75e6c571d6134"
+          }
+        },
+        "84e6829bb88845a8a4f42700b8496925": {
+          "model_module": "@jupyter-widgets/controls",
+          "model_name": "HTMLModel",
+          "model_module_version": "1.5.0",
+          "state": {
+            "_dom_classes": [],
+            "_model_module": "@jupyter-widgets/controls",
+            "_model_module_version": "1.5.0",
+            "_model_name": "HTMLModel",
+            "_view_count": null,
+            "_view_module": "@jupyter-widgets/controls",
+            "_view_module_version": "1.5.0",
+            "_view_name": "HTMLView",
+            "description": "",
+            "description_tooltip": null,
+            "layout": "IPY_MODEL_c66a77395e42424d904699edcbb67291",
+            "placeholder": "​",
+            "style": "IPY_MODEL_c4bbc15bf853439399dbcf1d40a5a407",
+            "value": "100%"
+          }
+        },
+        "c038e52d41bf4d5b9602930c3d074087": {
+          "model_module": "@jupyter-widgets/controls",
+          "model_name": "FloatProgressModel",
+          "model_module_version": "1.5.0",
+          "state": {
+            "_dom_classes": [],
+            "_model_module": "@jupyter-widgets/controls",
+            "_model_module_version": "1.5.0",
+            "_model_name": "FloatProgressModel",
+            "_view_count": null,
+            "_view_module": "@jupyter-widgets/controls",
+            "_view_module_version": "1.5.0",
+            "_view_name": "ProgressView",
+            "bar_style": "success",
+            "description": "",
+            "description_tooltip": null,
+            "layout": "IPY_MODEL_0aaabfac395b43afbdd6d752c502bbf6",
+            "max": 818322941,
+            "min": 0,
+            "orientation": "horizontal",
+            "style": "IPY_MODEL_3786d970492b4aa38f886f2572fd958c",
+            "value": 818322941
+          }
+        },
+        "2667604641764341b0bc8c6afea438fd": {
+          "model_module": "@jupyter-widgets/controls",
+          "model_name": "HTMLModel",
+          "model_module_version": "1.5.0",
+          "state": {
+            "_dom_classes": [],
+            "_model_module": "@jupyter-widgets/controls",
+            "_model_module_version": "1.5.0",
+            "_model_name": "HTMLModel",
+            "_view_count": null,
+            "_view_module": "@jupyter-widgets/controls",
+            "_view_module_version": "1.5.0",
+            "_view_name": "HTMLView",
+            "description": "",
+            "description_tooltip": null,
+            "layout": "IPY_MODEL_b86d0f2d7be74cebbcaa884b53123eeb",
+            "placeholder": "​",
+            "style": "IPY_MODEL_fa7b1497925a457f89286a71f073f416",
+            "value": " 780M/780M [00:57<00:00, 10.1MB/s]"
+          }
+        },
+        "98b3a4806ed14102b0d75e6c571d6134": {
+          "model_module": "@jupyter-widgets/base",
+          "model_name": "LayoutModel",
+          "model_module_version": "1.2.0",
+          "state": {
+            "_model_module": "@jupyter-widgets/base",
+            "_model_module_version": "1.2.0",
+            "_model_name": "LayoutModel",
+            "_view_count": null,
+            "_view_module": "@jupyter-widgets/base",
+            "_view_module_version": "1.2.0",
+            "_view_name": "LayoutView",
+            "align_content": null,
+            "align_items": null,
+            "align_self": null,
+            "border": null,
+            "bottom": null,
+            "display": null,
+            "flex": null,
+            "flex_flow": null,
+            "grid_area": null,
+            "grid_auto_columns": null,
+            "grid_auto_flow": null,
+            "grid_auto_rows": null,
+            "grid_column": null,
+            "grid_gap": null,
+            "grid_row": null,
+            "grid_template_areas": null,
+            "grid_template_columns": null,
+            "grid_template_rows": null,
+            "height": null,
+            "justify_content": null,
+            "justify_items": null,
+            "left": null,
+            "margin": null,
+            "max_height": null,
+            "max_width": null,
+            "min_height": null,
+            "min_width": null,
+            "object_fit": null,
+            "object_position": null,
+            "order": null,
+            "overflow": null,
+            "overflow_x": null,
+            "overflow_y": null,
+            "padding": null,
+            "right": null,
+            "top": null,
+            "visibility": null,
+            "width": null
+          }
+        },
+        "c66a77395e42424d904699edcbb67291": {
+          "model_module": "@jupyter-widgets/base",
+          "model_name": "LayoutModel",
+          "model_module_version": "1.2.0",
+          "state": {
+            "_model_module": "@jupyter-widgets/base",
+            "_model_module_version": "1.2.0",
+            "_model_name": "LayoutModel",
+            "_view_count": null,
+            "_view_module": "@jupyter-widgets/base",
+            "_view_module_version": "1.2.0",
+            "_view_name": "LayoutView",
+            "align_content": null,
+            "align_items": null,
+            "align_self": null,
+            "border": null,
+            "bottom": null,
+            "display": null,
+            "flex": null,
+            "flex_flow": null,
+            "grid_area": null,
+            "grid_auto_columns": null,
+            "grid_auto_flow": null,
+            "grid_auto_rows": null,
+            "grid_column": null,
+            "grid_gap": null,
+            "grid_row": null,
+            "grid_template_areas": null,
+            "grid_template_columns": null,
+            "grid_template_rows": null,
+            "height": null,
+            "justify_content": null,
+            "justify_items": null,
+            "left": null,
+            "margin": null,
+            "max_height": null,
+            "max_width": null,
+            "min_height": null,
+            "min_width": null,
+            "object_fit": null,
+            "object_position": null,
+            "order": null,
+            "overflow": null,
+            "overflow_x": null,
+            "overflow_y": null,
+            "padding": null,
+            "right": null,
+            "top": null,
+            "visibility": null,
+            "width": null
+          }
+        },
+        "c4bbc15bf853439399dbcf1d40a5a407": {
+          "model_module": "@jupyter-widgets/controls",
+          "model_name": "DescriptionStyleModel",
+          "model_module_version": "1.5.0",
+          "state": {
+            "_model_module": "@jupyter-widgets/controls",
+            "_model_module_version": "1.5.0",
+            "_model_name": "DescriptionStyleModel",
+            "_view_count": null,
+            "_view_module": "@jupyter-widgets/base",
+            "_view_module_version": "1.2.0",
+            "_view_name": "StyleView",
+            "description_width": ""
+          }
+        },
+        "0aaabfac395b43afbdd6d752c502bbf6": {
+          "model_module": "@jupyter-widgets/base",
+          "model_name": "LayoutModel",
+          "model_module_version": "1.2.0",
+          "state": {
+            "_model_module": "@jupyter-widgets/base",
+            "_model_module_version": "1.2.0",
+            "_model_name": "LayoutModel",
+            "_view_count": null,
+            "_view_module": "@jupyter-widgets/base",
+            "_view_module_version": "1.2.0",
+            "_view_name": "LayoutView",
+            "align_content": null,
+            "align_items": null,
+            "align_self": null,
+            "border": null,
+            "bottom": null,
+            "display": null,
+            "flex": null,
+            "flex_flow": null,
+            "grid_area": null,
+            "grid_auto_columns": null,
+            "grid_auto_flow": null,
+            "grid_auto_rows": null,
+            "grid_column": null,
+            "grid_gap": null,
+            "grid_row": null,
+            "grid_template_areas": null,
+            "grid_template_columns": null,
+            "grid_template_rows": null,
+            "height": null,
+            "justify_content": null,
+            "justify_items": null,
+            "left": null,
+            "margin": null,
+            "max_height": null,
+            "max_width": null,
+            "min_height": null,
+            "min_width": null,
+            "object_fit": null,
+            "object_position": null,
+            "order": null,
+            "overflow": null,
+            "overflow_x": null,
+            "overflow_y": null,
+            "padding": null,
+            "right": null,
+            "top": null,
+            "visibility": null,
+            "width": null
+          }
+        },
+        "3786d970492b4aa38f886f2572fd958c": {
+          "model_module": "@jupyter-widgets/controls",
+          "model_name": "ProgressStyleModel",
+          "model_module_version": "1.5.0",
+          "state": {
+            "_model_module": "@jupyter-widgets/controls",
+            "_model_module_version": "1.5.0",
+            "_model_name": "ProgressStyleModel",
+            "_view_count": null,
+            "_view_module": "@jupyter-widgets/base",
+            "_view_module_version": "1.2.0",
+            "_view_name": "StyleView",
+            "bar_color": null,
+            "description_width": ""
+          }
+        },
+        "b86d0f2d7be74cebbcaa884b53123eeb": {
+          "model_module": "@jupyter-widgets/base",
+          "model_name": "LayoutModel",
+          "model_module_version": "1.2.0",
+          "state": {
+            "_model_module": "@jupyter-widgets/base",
+            "_model_module_version": "1.2.0",
+            "_model_name": "LayoutModel",
+            "_view_count": null,
+            "_view_module": "@jupyter-widgets/base",
+            "_view_module_version": "1.2.0",
+            "_view_name": "LayoutView",
+            "align_content": null,
+            "align_items": null,
+            "align_self": null,
+            "border": null,
+            "bottom": null,
+            "display": null,
+            "flex": null,
+            "flex_flow": null,
+            "grid_area": null,
+            "grid_auto_columns": null,
+            "grid_auto_flow": null,
+            "grid_auto_rows": null,
+            "grid_column": null,
+            "grid_gap": null,
+            "grid_row": null,
+            "grid_template_areas": null,
+            "grid_template_columns": null,
+            "grid_template_rows": null,
+            "height": null,
+            "justify_content": null,
+            "justify_items": null,
+            "left": null,
+            "margin": null,
+            "max_height": null,
+            "max_width": null,
+            "min_height": null,
+            "min_width": null,
+            "object_fit": null,
+            "object_position": null,
+            "order": null,
+            "overflow": null,
+            "overflow_x": null,
+            "overflow_y": null,
+            "padding": null,
+            "right": null,
+            "top": null,
+            "visibility": null,
+            "width": null
+          }
+        },
+        "fa7b1497925a457f89286a71f073f416": {
+          "model_module": "@jupyter-widgets/controls",
+          "model_name": "DescriptionStyleModel",
+          "model_module_version": "1.5.0",
+          "state": {
+            "_model_module": "@jupyter-widgets/controls",
+            "_model_module_version": "1.5.0",
+            "_model_name": "DescriptionStyleModel",
+            "_view_count": null,
+            "_view_module": "@jupyter-widgets/base",
+            "_view_module_version": "1.2.0",
+            "_view_name": "StyleView",
+            "description_width": ""
+          }
+        }
+      }
+    }
+  },
+  "cells": [
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "t6MPjfT5NrKQ"
+      },
+      "source": [
+        "<div align=\"center\">\n",
+        "\n",
+        "  <a href=\"https://ultralytics.com/yolov5\" target=\"_blank\">\n",
+        "    <img width=\"1024\", src=\"https://raw.githubusercontent.com/ultralytics/assets/master/yolov5/v70/splash.png\"></a>\n",
+        "\n",
+        "\n",
+        "<br>\n",
+        "  <a href=\"https://bit.ly/yolov5-paperspace-notebook\"><img src=\"https://assets.paperspace.io/img/gradient-badge.svg\" alt=\"Run on Gradient\"></a>\n",
+        "  <a href=\"https://colab.research.google.com/github/ultralytics/yolov5/blob/master/tutorial.ipynb\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"></a>\n",
+        "  <a href=\"https://www.kaggle.com/ultralytics/yolov5\"><img src=\"https://kaggle.com/static/images/open-in-kaggle.svg\" alt=\"Open In Kaggle\"></a>\n",
+        "<br>\n",
+        "\n",
+        "This <a href=\"https://github.com/ultralytics/yolov5\">YOLOv5</a> 🚀 notebook by <a href=\"https://ultralytics.com\">Ultralytics</a> presents simple train, validate and predict examples to help start your AI adventure.<br>See <a href=\"https://github.com/ultralytics/yolov5/issues/new/choose\">GitHub</a> for community support or <a href=\"https://ultralytics.com/contact\">contact us</a> for professional support.\n",
+        "\n",
+        "</div>"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "7mGmQbAO5pQb"
+      },
+      "source": [
+        "# Setup\n",
+        "\n",
+        "Clone GitHub [repository](https://github.com/ultralytics/yolov5), install [dependencies](https://github.com/ultralytics/yolov5/blob/master/requirements.txt) and check PyTorch and GPU."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "wbvMlHd_QwMG",
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "outputId": "32e3bc15-6d02-4352-f0a3-912059d134a5"
+      },
+      "source": [
+        "!git clone https://github.com/ultralytics/yolov5  # clone\n",
+        "%cd yolov5\n",
+        "%pip install -qr requirements.txt  # install\n",
+        "\n",
+        "import torch\n",
+        "import utils\n",
+        "display = utils.notebook_init()  # checks"
+      ],
+      "execution_count": 1,
+      "outputs": [
+        {
+          "output_type": "stream",
+          "name": "stderr",
+          "text": [
+            "YOLOv5 🚀 v6.2-256-g0051615 Python-3.7.15 torch-1.12.1+cu113 CUDA:0 (Tesla T4, 15110MiB)\n"
+          ]
+        },
+        {
+          "output_type": "stream",
+          "name": "stdout",
+          "text": [
+            "Setup complete ✅ (2 CPUs, 12.7 GB RAM, 22.6/78.2 GB disk)\n"
+          ]
+        }
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "4JnkELT0cIJg"
+      },
+      "source": [
+        "# 1. Detect\n",
+        "\n",
+        "`detect.py` runs YOLOv5 inference on a variety of sources, downloading models automatically from the [latest YOLOv5 release](https://github.com/ultralytics/yolov5/releases), and saving results to `runs/detect`. Example inference sources are:\n",
+        "\n",
+        "```shell\n",
+        "python detect.py --source 0  # webcam\n",
+        "                          img.jpg  # image \n",
+        "                          vid.mp4  # video\n",
+        "                          screen  # screenshot\n",
+        "                          path/  # directory\n",
+        "                          'path/*.jpg'  # glob\n",
+        "                          'https://youtu.be/Zgi9g1ksQHc'  # YouTube\n",
+        "                          'rtsp://example.com/media.mp4'  # RTSP, RTMP, HTTP stream\n",
+        "```"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "zR9ZbuQCH7FX",
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "outputId": "8e81d6e9-0360-4212-cd61-9a5a58d3f703"
+      },
+      "source": [
+        "!python detect.py --weights yolov5s.pt --img 640 --conf 0.25 --source data/images\n",
+        "# display.Image(filename='runs/detect/exp/zidane.jpg', width=600)"
+      ],
+      "execution_count": 2,
+      "outputs": [
+        {
+          "output_type": "stream",
+          "name": "stdout",
+          "text": [
+            "\u001b[34m\u001b[1mdetect: \u001b[0mweights=['yolov5s.pt'], source=data/images, data=data/coco128.yaml, imgsz=[640, 640], conf_thres=0.25, iou_thres=0.45, max_det=1000, device=, view_img=False, save_txt=False, save_conf=False, save_crop=False, nosave=False, classes=None, agnostic_nms=False, augment=False, visualize=False, update=False, project=runs/detect, name=exp, exist_ok=False, line_thickness=3, hide_labels=False, hide_conf=False, half=False, dnn=False, vid_stride=1\n",
+            "YOLOv5 🚀 v6.2-256-g0051615 Python-3.7.15 torch-1.12.1+cu113 CUDA:0 (Tesla T4, 15110MiB)\n",
+            "\n",
+            "Downloading https://github.com/ultralytics/yolov5/releases/download/v6.2/yolov5s.pt to yolov5s.pt...\n",
+            "100% 14.1M/14.1M [00:00<00:00, 19.5MB/s]\n",
+            "\n",
+            "Fusing layers... \n",
+            "YOLOv5s summary: 213 layers, 7225885 parameters, 0 gradients\n",
+            "image 1/2 /content/yolov5/data/images/bus.jpg: 640x480 4 persons, 1 bus, 17.5ms\n",
+            "image 2/2 /content/yolov5/data/images/zidane.jpg: 384x640 2 persons, 2 ties, 18.0ms\n",
+            "Speed: 0.5ms pre-process, 17.8ms inference, 17.6ms NMS per image at shape (1, 3, 640, 640)\n",
+            "Results saved to \u001b[1mruns/detect/exp\u001b[0m\n"
+          ]
+        }
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "hkAzDWJ7cWTr"
+      },
+      "source": [
+        "&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;\n",
+        "<img align=\"left\" src=\"https://user-images.githubusercontent.com/26833433/127574988-6a558aa1-d268-44b9-bf6b-62d4c605cc72.jpg\" width=\"600\">"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "0eq1SMWl6Sfn"
+      },
+      "source": [
+        "# 2. Validate\n",
+        "Validate a model's accuracy on the [COCO](https://cocodataset.org/#home) dataset's `val` or `test` splits. Models are downloaded automatically from the [latest YOLOv5 release](https://github.com/ultralytics/yolov5/releases). To show results by class use the `--verbose` flag."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "WQPtK1QYVaD_",
+        "colab": {
+          "base_uri": "https://localhost:8080/",
+          "height": 49,
+          "referenced_widgets": [
+            "300b4d5355ef4967bd5246afeef6eef5",
+            "84e6829bb88845a8a4f42700b8496925",
+            "c038e52d41bf4d5b9602930c3d074087",
+            "2667604641764341b0bc8c6afea438fd",
+            "98b3a4806ed14102b0d75e6c571d6134",
+            "c66a77395e42424d904699edcbb67291",
+            "c4bbc15bf853439399dbcf1d40a5a407",
+            "0aaabfac395b43afbdd6d752c502bbf6",
+            "3786d970492b4aa38f886f2572fd958c",
+            "b86d0f2d7be74cebbcaa884b53123eeb",
+            "fa7b1497925a457f89286a71f073f416"
+          ]
+        },
+        "outputId": "61ffec5e-90ea-44f6-c0ea-b006e6e7072f"
+      },
+      "source": [
+        "# Download COCO val\n",
+        "torch.hub.download_url_to_file('https://ultralytics.com/assets/coco2017val.zip', 'tmp.zip')  # download (780M - 5000 images)\n",
+        "!unzip -q tmp.zip -d ../datasets && rm tmp.zip  # unzip"
+      ],
+      "execution_count": 3,
+      "outputs": [
+        {
+          "output_type": "display_data",
+          "data": {
+            "text/plain": [
+              "  0%|          | 0.00/780M [00:00<?, ?B/s]"
+            ],
+            "application/vnd.jupyter.widget-view+json": {
+              "version_major": 2,
+              "version_minor": 0,
+              "model_id": "300b4d5355ef4967bd5246afeef6eef5"
+            }
+          },
+          "metadata": {}
+        }
+      ]
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "X58w8JLpMnjH",
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "outputId": "aa5d5cea-14c1-4a19-bfdf-95b7164962cf"
+      },
+      "source": [
+        "# Validate YOLOv5s on COCO val\n",
+        "!python val.py --weights yolov5s.pt --data coco.yaml --img 640 --half"
+      ],
+      "execution_count": 4,
+      "outputs": [
+        {
+          "output_type": "stream",
+          "name": "stdout",
+          "text": [
+            "\u001b[34m\u001b[1mval: \u001b[0mdata=/content/yolov5/data/coco.yaml, weights=['yolov5s.pt'], batch_size=32, imgsz=640, conf_thres=0.001, iou_thres=0.6, max_det=300, task=val, device=, workers=8, single_cls=False, augment=False, verbose=False, save_txt=False, save_hybrid=False, save_conf=False, save_json=True, project=runs/val, name=exp, exist_ok=False, half=True, dnn=False\n",
+            "YOLOv5 🚀 v6.2-256-g0051615 Python-3.7.15 torch-1.12.1+cu113 CUDA:0 (Tesla T4, 15110MiB)\n",
+            "\n",
+            "Fusing layers... \n",
+            "YOLOv5s summary: 213 layers, 7225885 parameters, 0 gradients\n",
+            "\u001b[34m\u001b[1mval: \u001b[0mScanning /content/datasets/coco/val2017... 4952 images, 48 backgrounds, 0 corrupt: 100% 5000/5000 [00:02<00:00, 2066.57it/s]\n",
+            "\u001b[34m\u001b[1mval: \u001b[0mNew cache created: /content/datasets/coco/val2017.cache\n",
+            "                 Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 157/157 [01:09<00:00,  2.26it/s]\n",
+            "                   all       5000      36335       0.67      0.521      0.566      0.371\n",
+            "Speed: 0.1ms pre-process, 2.7ms inference, 1.9ms NMS per image at shape (32, 3, 640, 640)\n",
+            "\n",
+            "Evaluating pycocotools mAP... saving runs/val/exp/yolov5s_predictions.json...\n",
+            "loading annotations into memory...\n",
+            "Done (t=0.82s)\n",
+            "creating index...\n",
+            "index created!\n",
+            "Loading and preparing results...\n",
+            "DONE (t=5.49s)\n",
+            "creating index...\n",
+            "index created!\n",
+            "Running per image evaluation...\n",
+            "Evaluate annotation type *bbox*\n",
+            "DONE (t=74.26s).\n",
+            "Accumulating evaluation results...\n",
+            "DONE (t=13.46s).\n",
+            " Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.374\n",
+            " Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=100 ] = 0.572\n",
+            " Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=100 ] = 0.402\n",
+            " Average Precision  (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.211\n",
+            " Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.423\n",
+            " Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.489\n",
+            " Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=  1 ] = 0.311\n",
+            " Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 10 ] = 0.516\n",
+            " Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.566\n",
+            " Average Recall     (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.378\n",
+            " Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.625\n",
+            " Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.723\n",
+            "Results saved to \u001b[1mruns/val/exp\u001b[0m\n"
+          ]
+        }
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "ZY2VXXXu74w5"
+      },
+      "source": [
+        "# 3. Train\n",
+        "\n",
+        "<p align=\"\"><a href=\"https://roboflow.com/?ref=ultralytics\"><img width=\"1000\" src=\"https://github.com/ultralytics/assets/raw/master/im/integrations-loop.png\"/></a></p>\n",
+        "Close the active learning loop by sampling images from your inference conditions with the `roboflow` pip package\n",
+        "<br><br>\n",
+        "\n",
+        "Train a YOLOv5s model on the [COCO128](https://www.kaggle.com/ultralytics/coco128) dataset with `--data coco128.yaml`, starting from pretrained `--weights yolov5s.pt`, or from randomly initialized `--weights '' --cfg yolov5s.yaml`.\n",
+        "\n",
+        "- **Pretrained [Models](https://github.com/ultralytics/yolov5/tree/master/models)** are downloaded\n",
+        "automatically from the [latest YOLOv5 release](https://github.com/ultralytics/yolov5/releases)\n",
+        "- **[Datasets](https://github.com/ultralytics/yolov5/tree/master/data)** available for autodownload include: [COCO](https://github.com/ultralytics/yolov5/blob/master/data/coco.yaml), [COCO128](https://github.com/ultralytics/yolov5/blob/master/data/coco128.yaml), [VOC](https://github.com/ultralytics/yolov5/blob/master/data/VOC.yaml), [Argoverse](https://github.com/ultralytics/yolov5/blob/master/data/Argoverse.yaml), [VisDrone](https://github.com/ultralytics/yolov5/blob/master/data/VisDrone.yaml), [GlobalWheat](https://github.com/ultralytics/yolov5/blob/master/data/GlobalWheat2020.yaml), [xView](https://github.com/ultralytics/yolov5/blob/master/data/xView.yaml), [Objects365](https://github.com/ultralytics/yolov5/blob/master/data/Objects365.yaml), [SKU-110K](https://github.com/ultralytics/yolov5/blob/master/data/SKU-110K.yaml).\n",
+        "- **Training Results** are saved to `runs/train/` with incrementing run directories, i.e. `runs/train/exp2`, `runs/train/exp3` etc.\n",
+        "<br><br>\n",
+        "\n",
+        "A **Mosaic Dataloader** is used for training which combines 4 images into 1 mosaic.\n",
+        "\n",
+        "## Train on Custom Data with Roboflow 🌟 NEW\n",
+        "\n",
+        "[Roboflow](https://roboflow.com/?ref=ultralytics) enables you to easily **organize, label, and prepare** a high quality dataset with your own custom data. Roboflow also makes it easy to establish an active learning pipeline, collaborate with your team on dataset improvement, and integrate directly into your model building workflow with the `roboflow` pip package.\n",
+        "\n",
+        "- Custom Training Example: [https://blog.roboflow.com/how-to-train-yolov5-on-a-custom-dataset/](https://blog.roboflow.com/how-to-train-yolov5-on-a-custom-dataset/?ref=ultralytics)\n",
+        "- Custom Training Notebook: [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/roboflow-ai/yolov5-custom-training-tutorial/blob/main/yolov5-custom-training.ipynb)\n",
+        "<br>\n",
+        "\n",
+        "<p align=\"\"><a href=\"https://roboflow.com/?ref=ultralytics\"><img width=\"480\" src=\"https://uploads-ssl.webflow.com/5f6bc60e665f54545a1e52a5/6152a275ad4b4ac20cd2e21a_roboflow-annotate.gif\"/></a></p>Label images lightning fast (including with model-assisted labeling)"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "source": [
+        "#@title Select YOLOv5 🚀 logger {run: 'auto'}\n",
+        "logger = 'TensorBoard' #@param ['TensorBoard', 'Comet', 'ClearML']\n",
+        "\n",
+        "if logger == 'TensorBoard':\n",
+        "  %load_ext tensorboard\n",
+        "  %tensorboard --logdir runs/train\n",
+        "elif logger == 'Comet':\n",
+        "  %pip install -q comet_ml\n",
+        "  import comet_ml; comet_ml.init()\n",
+        "elif logger == 'ClearML':\n",
+        "  %pip install -q clearml\n",
+        "  import clearml; clearml.browser_login()"
+      ],
+      "metadata": {
+        "id": "i3oKtE4g-aNn"
+      },
+      "execution_count": null,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "1NcFxRcFdJ_O",
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "outputId": "f0fcdc77-5326-41e1-bacc-be5432eefa2a"
+      },
+      "source": [
+        "# Train YOLOv5s on COCO128 for 3 epochs\n",
+        "!python train.py --img 640 --batch 16 --epochs 3 --data coco128.yaml --weights yolov5s.pt --cache"
+      ],
+      "execution_count": 5,
+      "outputs": [
+        {
+          "output_type": "stream",
+          "name": "stdout",
+          "text": [
+            "\u001b[34m\u001b[1mtrain: \u001b[0mweights=yolov5s.pt, cfg=, data=coco128.yaml, hyp=data/hyps/hyp.scratch-low.yaml, epochs=3, batch_size=16, imgsz=640, rect=False, resume=False, nosave=False, noval=False, noautoanchor=False, noplots=False, evolve=None, bucket=, cache=ram, image_weights=False, device=, multi_scale=False, single_cls=False, optimizer=SGD, sync_bn=False, workers=8, project=runs/train, name=exp, exist_ok=False, quad=False, cos_lr=False, label_smoothing=0.0, patience=100, freeze=[0], save_period=-1, seed=0, local_rank=-1, entity=None, upload_dataset=False, bbox_interval=-1, artifact_alias=latest\n",
+            "\u001b[34m\u001b[1mgithub: \u001b[0mup to date with https://github.com/ultralytics/yolov5 ✅\n",
+            "YOLOv5 🚀 v6.2-256-g0051615 Python-3.7.15 torch-1.12.1+cu113 CUDA:0 (Tesla T4, 15110MiB)\n",
+            "\n",
+            "\u001b[34m\u001b[1mhyperparameters: \u001b[0mlr0=0.01, lrf=0.01, momentum=0.937, weight_decay=0.0005, warmup_epochs=3.0, warmup_momentum=0.8, warmup_bias_lr=0.1, box=0.05, cls=0.5, cls_pw=1.0, obj=1.0, obj_pw=1.0, iou_t=0.2, anchor_t=4.0, fl_gamma=0.0, hsv_h=0.015, hsv_s=0.7, hsv_v=0.4, degrees=0.0, translate=0.1, scale=0.5, shear=0.0, perspective=0.0, flipud=0.0, fliplr=0.5, mosaic=1.0, mixup=0.0, copy_paste=0.0\n",
+            "\u001b[34m\u001b[1mClearML: \u001b[0mrun 'pip install clearml' to automatically track, visualize and remotely train YOLOv5 🚀 in ClearML\n",
+            "\u001b[34m\u001b[1mComet: \u001b[0mrun 'pip install comet_ml' to automatically track and visualize YOLOv5 🚀 runs in Comet\n",
+            "\u001b[34m\u001b[1mTensorBoard: \u001b[0mStart with 'tensorboard --logdir runs/train', view at http://localhost:6006/\n",
+            "\n",
+            "Dataset not found ⚠️, missing paths ['/content/datasets/coco128/images/train2017']\n",
+            "Downloading https://ultralytics.com/assets/coco128.zip to coco128.zip...\n",
+            "100% 6.66M/6.66M [00:00<00:00, 39.8MB/s]\n",
+            "Dataset download success ✅ (0.8s), saved to \u001b[1m/content/datasets\u001b[0m\n",
+            "\n",
+            "                 from  n    params  module                                  arguments                     \n",
+            "  0                -1  1      3520  models.common.Conv                      [3, 32, 6, 2, 2]              \n",
+            "  1                -1  1     18560  models.common.Conv                      [32, 64, 3, 2]                \n",
+            "  2                -1  1     18816  models.common.C3                        [64, 64, 1]                   \n",
+            "  3                -1  1     73984  models.common.Conv                      [64, 128, 3, 2]               \n",
+            "  4                -1  2    115712  models.common.C3                        [128, 128, 2]                 \n",
+            "  5                -1  1    295424  models.common.Conv                      [128, 256, 3, 2]              \n",
+            "  6                -1  3    625152  models.common.C3                        [256, 256, 3]                 \n",
+            "  7                -1  1   1180672  models.common.Conv                      [256, 512, 3, 2]              \n",
+            "  8                -1  1   1182720  models.common.C3                        [512, 512, 1]                 \n",
+            "  9                -1  1    656896  models.common.SPPF                      [512, 512, 5]                 \n",
+            " 10                -1  1    131584  models.common.Conv                      [512, 256, 1, 1]              \n",
+            " 11                -1  1         0  torch.nn.modules.upsampling.Upsample    [None, 2, 'nearest']          \n",
+            " 12           [-1, 6]  1         0  models.common.Concat                    [1]                           \n",
+            " 13                -1  1    361984  models.common.C3                        [512, 256, 1, False]          \n",
+            " 14                -1  1     33024  models.common.Conv                      [256, 128, 1, 1]              \n",
+            " 15                -1  1         0  torch.nn.modules.upsampling.Upsample    [None, 2, 'nearest']          \n",
+            " 16           [-1, 4]  1         0  models.common.Concat                    [1]                           \n",
+            " 17                -1  1     90880  models.common.C3                        [256, 128, 1, False]          \n",
+            " 18                -1  1    147712  models.common.Conv                      [128, 128, 3, 2]              \n",
+            " 19          [-1, 14]  1         0  models.common.Concat                    [1]                           \n",
+            " 20                -1  1    296448  models.common.C3                        [256, 256, 1, False]          \n",
+            " 21                -1  1    590336  models.common.Conv                      [256, 256, 3, 2]              \n",
+            " 22          [-1, 10]  1         0  models.common.Concat                    [1]                           \n",
+            " 23                -1  1   1182720  models.common.C3                        [512, 512, 1, False]          \n",
+            " 24      [17, 20, 23]  1    229245  models.yolo.Detect                      [80, [[10, 13, 16, 30, 33, 23], [30, 61, 62, 45, 59, 119], [116, 90, 156, 198, 373, 326]], [128, 256, 512]]\n",
+            "Model summary: 214 layers, 7235389 parameters, 7235389 gradients, 16.6 GFLOPs\n",
+            "\n",
+            "Transferred 349/349 items from yolov5s.pt\n",
+            "\u001b[34m\u001b[1mAMP: \u001b[0mchecks passed ✅\n",
+            "\u001b[34m\u001b[1moptimizer:\u001b[0m SGD(lr=0.01) with parameter groups 57 weight(decay=0.0), 60 weight(decay=0.0005), 60 bias\n",
+            "\u001b[34m\u001b[1malbumentations: \u001b[0mBlur(p=0.01, blur_limit=(3, 7)), MedianBlur(p=0.01, blur_limit=(3, 7)), ToGray(p=0.01), CLAHE(p=0.01, clip_limit=(1, 4.0), tile_grid_size=(8, 8))\n",
+            "\u001b[34m\u001b[1mtrain: \u001b[0mScanning /content/datasets/coco128/labels/train2017... 126 images, 2 backgrounds, 0 corrupt: 100% 128/128 [00:00<00:00, 2084.63it/s]\n",
+            "\u001b[34m\u001b[1mtrain: \u001b[0mNew cache created: /content/datasets/coco128/labels/train2017.cache\n",
+            "\u001b[34m\u001b[1mtrain: \u001b[0mCaching images (0.1GB ram): 100% 128/128 [00:00<00:00, 255.09it/s]\n",
+            "\u001b[34m\u001b[1mval: \u001b[0mScanning /content/datasets/coco128/labels/train2017.cache... 126 images, 2 backgrounds, 0 corrupt: 100% 128/128 [00:00<?, ?it/s]\n",
+            "\u001b[34m\u001b[1mval: \u001b[0mCaching images (0.1GB ram): 100% 128/128 [00:01<00:00, 106.58it/s]\n",
+            "\n",
+            "\u001b[34m\u001b[1mAutoAnchor: \u001b[0m4.27 anchors/target, 0.994 Best Possible Recall (BPR). Current anchors are a good fit to dataset ✅\n",
+            "Plotting labels to runs/train/exp/labels.jpg... \n",
+            "Image sizes 640 train, 640 val\n",
+            "Using 2 dataloader workers\n",
+            "Logging results to \u001b[1mruns/train/exp\u001b[0m\n",
+            "Starting training for 3 epochs...\n",
+            "\n",
+            "      Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size\n",
+            "        0/2      3.74G    0.04618    0.07207      0.017        232        640: 100% 8/8 [00:06<00:00,  1.33it/s]\n",
+            "                 Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 4/4 [00:01<00:00,  2.99it/s]\n",
+            "                   all        128        929      0.672      0.594      0.682      0.451\n",
+            "\n",
+            "      Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size\n",
+            "        1/2      5.36G    0.04623    0.06888    0.01821        201        640: 100% 8/8 [00:02<00:00,  3.28it/s]\n",
+            "                 Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 4/4 [00:01<00:00,  3.02it/s]\n",
+            "                   all        128        929      0.721      0.639      0.724       0.48\n",
+            "\n",
+            "      Epoch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size\n",
+            "        2/2      5.36G    0.04361    0.06479    0.01698        227        640: 100% 8/8 [00:02<00:00,  3.50it/s]\n",
+            "                 Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 4/4 [00:01<00:00,  3.05it/s]\n",
+            "                   all        128        929      0.758      0.641      0.731      0.487\n",
+            "\n",
+            "3 epochs completed in 0.005 hours.\n",
+            "Optimizer stripped from runs/train/exp/weights/last.pt, 14.9MB\n",
+            "Optimizer stripped from runs/train/exp/weights/best.pt, 14.9MB\n",
+            "\n",
+            "Validating runs/train/exp/weights/best.pt...\n",
+            "Fusing layers... \n",
+            "Model summary: 157 layers, 7225885 parameters, 0 gradients, 16.4 GFLOPs\n",
+            "                 Class     Images  Instances          P          R      mAP50   mAP50-95: 100% 4/4 [00:03<00:00,  1.09it/s]\n",
+            "                   all        128        929      0.757      0.641      0.732      0.487\n",
+            "                person        128        254       0.86      0.705      0.804      0.528\n",
+            "               bicycle        128          6      0.773      0.578      0.725      0.426\n",
+            "                   car        128         46      0.658      0.435      0.554      0.239\n",
+            "            motorcycle        128          5       0.59        0.8      0.837      0.635\n",
+            "              airplane        128          6          1      0.996      0.995      0.696\n",
+            "                   bus        128          7      0.635      0.714      0.756      0.666\n",
+            "                 train        128          3      0.691      0.333      0.753      0.511\n",
+            "                 truck        128         12      0.604      0.333      0.472       0.26\n",
+            "                  boat        128          6      0.941      0.333       0.46      0.183\n",
+            "         traffic light        128         14      0.557      0.183      0.302      0.214\n",
+            "             stop sign        128          2      0.827          1      0.995      0.846\n",
+            "                 bench        128          9       0.79      0.556      0.677      0.318\n",
+            "                  bird        128         16      0.962          1      0.995      0.663\n",
+            "                   cat        128          4      0.867          1      0.995      0.754\n",
+            "                   dog        128          9          1      0.649      0.903      0.654\n",
+            "                 horse        128          2      0.853          1      0.995      0.622\n",
+            "              elephant        128         17      0.908      0.882      0.934      0.698\n",
+            "                  bear        128          1      0.697          1      0.995      0.995\n",
+            "                 zebra        128          4      0.867          1      0.995      0.905\n",
+            "               giraffe        128          9      0.788      0.829      0.912      0.701\n",
+            "              backpack        128          6      0.841        0.5      0.738      0.311\n",
+            "              umbrella        128         18      0.786      0.815      0.859       0.48\n",
+            "               handbag        128         19      0.772      0.263      0.366      0.216\n",
+            "                   tie        128          7      0.975      0.714       0.77      0.491\n",
+            "              suitcase        128          4      0.643       0.75      0.912      0.563\n",
+            "               frisbee        128          5       0.72        0.8       0.76      0.717\n",
+            "                  skis        128          1      0.748          1      0.995        0.3\n",
+            "             snowboard        128          7      0.827      0.686      0.833       0.57\n",
+            "           sports ball        128          6      0.637      0.667      0.602      0.311\n",
+            "                  kite        128         10      0.645        0.6      0.594      0.224\n",
+            "          baseball bat        128          4      0.519      0.278      0.468      0.205\n",
+            "        baseball glove        128          7      0.483      0.429      0.465      0.278\n",
+            "            skateboard        128          5      0.923        0.6      0.687      0.493\n",
+            "         tennis racket        128          7      0.774      0.429      0.544      0.333\n",
+            "                bottle        128         18      0.577      0.379      0.551      0.275\n",
+            "            wine glass        128         16      0.715      0.875      0.893      0.511\n",
+            "                   cup        128         36      0.843      0.667      0.833      0.531\n",
+            "                  fork        128          6      0.998      0.333       0.45      0.315\n",
+            "                 knife        128         16       0.77      0.688      0.695      0.399\n",
+            "                 spoon        128         22      0.839      0.473      0.638      0.383\n",
+            "                  bowl        128         28      0.765      0.583      0.715      0.512\n",
+            "                banana        128          1      0.903          1      0.995      0.301\n",
+            "              sandwich        128          2          1          0      0.359      0.301\n",
+            "                orange        128          4      0.718       0.75      0.912      0.581\n",
+            "              broccoli        128         11      0.545      0.364       0.43      0.319\n",
+            "                carrot        128         24       0.62      0.625      0.724      0.495\n",
+            "               hot dog        128          2      0.385          1      0.828      0.762\n",
+            "                 pizza        128          5      0.833          1      0.962      0.725\n",
+            "                 donut        128         14      0.631          1       0.96      0.833\n",
+            "                  cake        128          4      0.871          1      0.995       0.83\n",
+            "                 chair        128         35      0.583        0.6      0.608      0.318\n",
+            "                 couch        128          6      0.909      0.667      0.813      0.543\n",
+            "          potted plant        128         14      0.745      0.786      0.822       0.48\n",
+            "                   bed        128          3      0.973      0.333      0.753       0.41\n",
+            "          dining table        128         13      0.821      0.356      0.577      0.342\n",
+            "                toilet        128          2          1      0.949      0.995      0.797\n",
+            "                    tv        128          2      0.566          1      0.995      0.796\n",
+            "                laptop        128          3          1          0       0.59      0.311\n",
+            "                 mouse        128          2          1          0      0.105     0.0527\n",
+            "                remote        128          8          1      0.623      0.634      0.538\n",
+            "            cell phone        128          8      0.565      0.375      0.399      0.179\n",
+            "             microwave        128          3      0.709          1      0.995      0.736\n",
+            "                  oven        128          5      0.328        0.4       0.43      0.282\n",
+            "                  sink        128          6      0.438      0.333      0.339      0.266\n",
+            "          refrigerator        128          5      0.564        0.8      0.798      0.535\n",
+            "                  book        128         29      0.597      0.256      0.351      0.155\n",
+            "                 clock        128          9      0.763      0.889      0.934      0.737\n",
+            "                  vase        128          2      0.331          1      0.995      0.895\n",
+            "              scissors        128          1          1          0      0.497     0.0552\n",
+            "            teddy bear        128         21      0.857       0.57      0.837      0.544\n",
+            "            toothbrush        128          5      0.799          1      0.928      0.556\n",
+            "Results saved to \u001b[1mruns/train/exp\u001b[0m\n"
+          ]
+        }
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "15glLzbQx5u0"
+      },
+      "source": [
+        "# 4. Visualize"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "source": [
+        "## Comet Logging and Visualization 🌟 NEW\n",
+        "[Comet](https://bit.ly/yolov5-readme-comet) is now fully integrated with YOLOv5. Track and visualize model metrics in real time, save your hyperparameters, datasets, and model checkpoints, and visualize your model predictions with [Comet Custom Panels](https://bit.ly/yolov5-colab-comet-panels)! Comet makes sure you never lose track of your work and makes it easy to share results and collaborate across teams of all sizes! \n",
+        "\n",
+        "Getting started is easy:\n",
+        "```shell\n",
+        "pip install comet_ml  # 1. install\n",
+        "export COMET_API_KEY=<Your API Key>  # 2. paste API key\n",
+        "python train.py --img 640 --epochs 3 --data coco128.yaml --weights yolov5s.pt  # 3. train\n",
+        "```\n",
+        "\n",
+        "To learn more about all of the supported Comet features for this integration, check out the [Comet Tutorial](https://github.com/ultralytics/yolov5/tree/master/utils/loggers/comet). If you'd like to learn more about Comet, head over to our [documentation](https://bit.ly/yolov5-colab-comet-docs). Get started by trying out the Comet Colab Notebook:\n",
+        "[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/drive/1RG0WOQyxlDlo5Km8GogJpIEJlg_5lyYO?usp=sharing)\n",
+        "\n",
+        "<img width=\"1920\" alt=\"yolo-ui\" src=\"https://user-images.githubusercontent.com/26833433/202851203-164e94e1-2238-46dd-91f8-de020e9d6b41.png\">"
+      ],
+      "metadata": {
+        "id": "nWOsI5wJR1o3"
+      }
+    },
+    {
+      "cell_type": "markdown",
+      "source": [
+        "## ClearML Logging and Automation 🌟 NEW\n",
+        "\n",
+        "[ClearML](https://cutt.ly/yolov5-notebook-clearml) is completely integrated into YOLOv5 to track your experimentation, manage dataset versions and even remotely execute training runs. To enable ClearML (check cells above):\n",
+        "\n",
+        "- `pip install clearml`\n",
+        "- run `clearml-init` to connect to a ClearML server (**deploy your own [open-source server](https://github.com/allegroai/clearml-server)**, or use our [free hosted server](https://cutt.ly/yolov5-notebook-clearml))\n",
+        "\n",
+        "You'll get all the great expected features from an experiment manager: live updates, model upload, experiment comparison etc. but ClearML also tracks uncommitted changes and installed packages for example. Thanks to that ClearML Tasks (which is what we call experiments) are also reproducible on different machines! With only 1 extra line, we can schedule a YOLOv5 training task on a queue to be executed by any number of ClearML Agents (workers).\n",
+        "\n",
+        "You can use ClearML Data to version your dataset and then pass it to YOLOv5 simply using its unique ID. This will help you keep track of your data without adding extra hassle. Explore the [ClearML Tutorial](https://github.com/ultralytics/yolov5/tree/master/utils/loggers/clearml) for details!\n",
+        "\n",
+        "<a href=\"https://cutt.ly/yolov5-notebook-clearml\">\n",
+        "<img alt=\"ClearML Experiment Management UI\" src=\"https://github.com/thepycoder/clearml_screenshots/raw/main/scalars.jpg\" width=\"1280\"/></a>"
+      ],
+      "metadata": {
+        "id": "Lay2WsTjNJzP"
+      }
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "-WPvRbS5Swl6"
+      },
+      "source": [
+        "## Local Logging\n",
+        "\n",
+        "Training results are automatically logged with [Tensorboard](https://www.tensorflow.org/tensorboard) and [CSV](https://github.com/ultralytics/yolov5/pull/4148) loggers to `runs/train`, with a new experiment directory created for each new training as `runs/train/exp2`, `runs/train/exp3`, etc.\n",
+        "\n",
+        "This directory contains train and val statistics, mosaics, labels, predictions and augmentated mosaics, as well as metrics and charts including precision-recall (PR) curves and confusion matrices. \n",
+        "\n",
+        "<img alt=\"Local logging results\" src=\"https://user-images.githubusercontent.com/26833433/183222430-e1abd1b7-782c-4cde-b04d-ad52926bf818.jpg\" width=\"1280\"/>\n"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "Zelyeqbyt3GD"
+      },
+      "source": [
+        "# Environments\n",
+        "\n",
+        "YOLOv5 may be run in any of the following up-to-date verified environments (with all dependencies including [CUDA](https://developer.nvidia.com/cuda)/[CUDNN](https://developer.nvidia.com/cudnn), [Python](https://www.python.org/) and [PyTorch](https://pytorch.org/) preinstalled):\n",
+        "\n",
+        "- **Notebooks** with free GPU: <a href=\"https://bit.ly/yolov5-paperspace-notebook\"><img src=\"https://assets.paperspace.io/img/gradient-badge.svg\" alt=\"Run on Gradient\"></a> <a href=\"https://colab.research.google.com/github/ultralytics/yolov5/blob/master/tutorial.ipynb\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"></a> <a href=\"https://www.kaggle.com/ultralytics/yolov5\"><img src=\"https://kaggle.com/static/images/open-in-kaggle.svg\" alt=\"Open In Kaggle\"></a>\n",
+        "- **Google Cloud** Deep Learning VM. See [GCP Quickstart Guide](https://github.com/ultralytics/yolov5/wiki/GCP-Quickstart)\n",
+        "- **Amazon** Deep Learning AMI. See [AWS Quickstart Guide](https://github.com/ultralytics/yolov5/wiki/AWS-Quickstart)\n",
+        "- **Docker Image**. See [Docker Quickstart Guide](https://github.com/ultralytics/yolov5/wiki/Docker-Quickstart) <a href=\"https://hub.docker.com/r/ultralytics/yolov5\"><img src=\"https://img.shields.io/docker/pulls/ultralytics/yolov5?logo=docker\" alt=\"Docker Pulls\"></a>\n"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "6Qu7Iesl0p54"
+      },
+      "source": [
+        "# Status\n",
+        "\n",
+        "![YOLOv5 CI](https://github.com/ultralytics/yolov5/actions/workflows/ci-testing.yml/badge.svg)\n",
+        "\n",
+        "If this badge is green, all [YOLOv5 GitHub Actions](https://github.com/ultralytics/yolov5/actions) Continuous Integration (CI) tests are currently passing. CI tests verify correct operation of YOLOv5 training ([train.py](https://github.com/ultralytics/yolov5/blob/master/train.py)), testing ([val.py](https://github.com/ultralytics/yolov5/blob/master/val.py)), inference ([detect.py](https://github.com/ultralytics/yolov5/blob/master/detect.py)) and export ([export.py](https://github.com/ultralytics/yolov5/blob/master/export.py)) on macOS, Windows, and Ubuntu every 24 hours and on every commit.\n"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "IEijrePND_2I"
+      },
+      "source": [
+        "# Appendix\n",
+        "\n",
+        "Additional content below."
+      ]
+    },
+    {
+      "cell_type": "code",
+      "metadata": {
+        "id": "GMusP4OAxFu6"
+      },
+      "source": [
+        "# YOLOv5 PyTorch HUB Inference (DetectionModels only)\n",
+        "import torch\n",
+        "\n",
+        "model = torch.hub.load('ultralytics/yolov5', 'yolov5s')  # yolov5n - yolov5x6 or custom\n",
+        "im = 'https://ultralytics.com/images/zidane.jpg'  # file, Path, PIL.Image, OpenCV, nparray, list\n",
+        "results = model(im)  # inference\n",
+        "results.print()  # or .show(), .save(), .crop(), .pandas(), etc."
+      ],
+      "execution_count": null,
+      "outputs": []
+    }
+  ]
+}

utils/__init__.py

+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+"""
+utils/initialization
+"""
+
+import contextlib
+import platform
+import threading
+
+
+def emojis(str=''):
+    # Return platform-dependent emoji-safe version of string
+    return str.encode().decode('ascii', 'ignore') if platform.system() == 'Windows' else str
+
+
+class TryExcept(contextlib.ContextDecorator):
+    # YOLOv5 TryExcept class. Usage: @TryExcept() decorator or 'with TryExcept():' context manager
+    def __init__(self, msg=''):
+        self.msg = msg
+
+    def __enter__(self):
+        pass
+
+    def __exit__(self, exc_type, value, traceback):
+        if value:
+            print(emojis(f"{self.msg}{': ' if self.msg else ''}{value}"))
+        return True
+
+
+def threaded(func):
+    # Multi-threads a target function and returns thread. Usage: @threaded decorator
+    def wrapper(*args, **kwargs):
+        thread = threading.Thread(target=func, args=args, kwargs=kwargs, daemon=True)
+        thread.start()
+        return thread
+
+    return wrapper
+
+
+def join_threads(verbose=False):
+    # Join all daemon threads, i.e. atexit.register(lambda: join_threads())
+    main_thread = threading.current_thread()
+    for t in threading.enumerate():
+        if t is not main_thread:
+            if verbose:
+                print(f'Joining thread {t.name}')
+            t.join()
+
+
+def notebook_init(verbose=True):
+    # Check system software and hardware
+    print('Checking setup...')
+
+    import os
+    import shutil
+
+    from utils.general import check_font, check_requirements, is_colab
+    from utils.torch_utils import select_device  # imports
+
+    check_font()
+
+    import psutil
+    from IPython import display  # to display images and clear console output
+
+    if is_colab():
+        shutil.rmtree('/content/sample_data', ignore_errors=True)  # remove colab /sample_data directory
+
+    # System info
+    if verbose:
+        gb = 1 << 30  # bytes to GiB (1024 ** 3)
+        ram = psutil.virtual_memory().total
+        total, used, free = shutil.disk_usage("/")
+        display.clear_output()
+        s = f'({os.cpu_count()} CPUs, {ram / gb:.1f} GB RAM, {(total - free) / gb:.1f}/{total / gb:.1f} GB disk)'
+    else:
+        s = ''
+
+    select_device(newline=False)
+    print(emojis(f'Setup complete ✅ {s}'))
+    return display

utils/activations.py

+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+"""
+Activation functions
+"""
+
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+
+
+class SiLU(nn.Module):
+    # SiLU activation https://arxiv.org/pdf/1606.08415.pdf
+    @staticmethod
+    def forward(x):
+        return x * torch.sigmoid(x)
+
+
+class Hardswish(nn.Module):
+    # Hard-SiLU activation
+    @staticmethod
+    def forward(x):
+        # return x * F.hardsigmoid(x)  # for TorchScript and CoreML
+        return x * F.hardtanh(x + 3, 0.0, 6.0) / 6.0  # for TorchScript, CoreML and ONNX
+
+
+class Mish(nn.Module):
+    # Mish activation https://github.com/digantamisra98/Mish
+    @staticmethod
+    def forward(x):
+        return x * F.softplus(x).tanh()
+
+
+class MemoryEfficientMish(nn.Module):
+    # Mish activation memory-efficient
+    class F(torch.autograd.Function):
+
+        @staticmethod
+        def forward(ctx, x):
+            ctx.save_for_backward(x)
+            return x.mul(torch.tanh(F.softplus(x)))  # x * tanh(ln(1 + exp(x)))
+
+        @staticmethod
+        def backward(ctx, grad_output):
+            x = ctx.saved_tensors[0]
+            sx = torch.sigmoid(x)
+            fx = F.softplus(x).tanh()
+            return grad_output * (fx + x * sx * (1 - fx * fx))
+
+    def forward(self, x):
+        return self.F.apply(x)
+
+
+class FReLU(nn.Module):
+    # FReLU activation https://arxiv.org/abs/2007.11824
+    def __init__(self, c1, k=3):  # ch_in, kernel
+        super().__init__()
+        self.conv = nn.Conv2d(c1, c1, k, 1, 1, groups=c1, bias=False)
+        self.bn = nn.BatchNorm2d(c1)
+
+    def forward(self, x):
+        return torch.max(x, self.bn(self.conv(x)))
+
+
+class AconC(nn.Module):
+    r""" ACON activation (activate or not)
+    AconC: (p1*x-p2*x) * sigmoid(beta*(p1*x-p2*x)) + p2*x, beta is a learnable parameter
+    according to "Activate or Not: Learning Customized Activation" <https://arxiv.org/pdf/2009.04759.pdf>.
+    """
+
+    def __init__(self, c1):
+        super().__init__()
+        self.p1 = nn.Parameter(torch.randn(1, c1, 1, 1))
+        self.p2 = nn.Parameter(torch.randn(1, c1, 1, 1))
+        self.beta = nn.Parameter(torch.ones(1, c1, 1, 1))
+
+    def forward(self, x):
+        dpx = (self.p1 - self.p2) * x
+        return dpx * torch.sigmoid(self.beta * dpx) + self.p2 * x
+
+
+class MetaAconC(nn.Module):
+    r""" ACON activation (activate or not)
+    MetaAconC: (p1*x-p2*x) * sigmoid(beta*(p1*x-p2*x)) + p2*x, beta is generated by a small network
+    according to "Activate or Not: Learning Customized Activation" <https://arxiv.org/pdf/2009.04759.pdf>.
+    """
+
+    def __init__(self, c1, k=1, s=1, r=16):  # ch_in, kernel, stride, r
+        super().__init__()
+        c2 = max(r, c1 // r)
+        self.p1 = nn.Parameter(torch.randn(1, c1, 1, 1))
+        self.p2 = nn.Parameter(torch.randn(1, c1, 1, 1))
+        self.fc1 = nn.Conv2d(c1, c2, k, s, bias=True)
+        self.fc2 = nn.Conv2d(c2, c1, k, s, bias=True)
+        # self.bn1 = nn.BatchNorm2d(c2)
+        # self.bn2 = nn.BatchNorm2d(c1)
+
+    def forward(self, x):
+        y = x.mean(dim=2, keepdims=True).mean(dim=3, keepdims=True)
+        # batch-size 1 bug/instabilities https://github.com/ultralytics/yolov5/issues/2891
+        # beta = torch.sigmoid(self.bn2(self.fc2(self.bn1(self.fc1(y)))))  # bug/unstable
+        beta = torch.sigmoid(self.fc2(self.fc1(y)))  # bug patch BN layers removed
+        dpx = (self.p1 - self.p2) * x
+        return dpx * torch.sigmoid(beta * dpx) + self.p2 * x

utils/augmentations.py

+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+"""
+Image augmentation functions
+"""
+
+import math
+import random
+
+import cv2
+import numpy as np
+import torch
+import torchvision.transforms as T
+import torchvision.transforms.functional as TF
+
+from utils.general import LOGGER, check_version, colorstr, resample_segments, segment2box, xywhn2xyxy
+from utils.metrics import bbox_ioa
+
+IMAGENET_MEAN = 0.485, 0.456, 0.406  # RGB mean
+IMAGENET_STD = 0.229, 0.224, 0.225  # RGB standard deviation
+
+
+class Albumentations:
+    # YOLOv5 Albumentations class (optional, only used if package is installed)
+    def __init__(self, size=640):
+        self.transform = None
+        prefix = colorstr('albumentations: ')
+        try:
+            import albumentations as A
+            check_version(A.__version__, '1.0.3', hard=True)  # version requirement
+
+            T = [
+                A.RandomResizedCrop(height=size, width=size, scale=(0.8, 1.0), ratio=(0.9, 1.11), p=0.0),
+                A.Blur(p=0.01),
+                A.MedianBlur(p=0.01),
+                A.ToGray(p=0.01),
+                A.CLAHE(p=0.01),
+                A.RandomBrightnessContrast(p=0.0),
+                A.RandomGamma(p=0.0),
+                A.ImageCompression(quality_lower=75, p=0.0)]  # transforms
+            self.transform = A.Compose(T, bbox_params=A.BboxParams(format='yolo', label_fields=['class_labels']))
+
+            LOGGER.info(prefix + ', '.join(f'{x}'.replace('always_apply=False, ', '') for x in T if x.p))
+        except ImportError:  # package not installed, skip
+            pass
+        except Exception as e:
+            LOGGER.info(f'{prefix}{e}')
+
+    def __call__(self, im, labels, p=1.0):
+        if self.transform and random.random() < p:
+            new = self.transform(image=im, bboxes=labels[:, 1:], class_labels=labels[:, 0])  # transformed
+            im, labels = new['image'], np.array([[c, *b] for c, b in zip(new['class_labels'], new['bboxes'])])
+        return im, labels
+
+
+def normalize(x, mean=IMAGENET_MEAN, std=IMAGENET_STD, inplace=False):
+    # Denormalize RGB images x per ImageNet stats in BCHW format, i.e. = (x - mean) / std
+    return TF.normalize(x, mean, std, inplace=inplace)
+
+
+def denormalize(x, mean=IMAGENET_MEAN, std=IMAGENET_STD):
+    # Denormalize RGB images x per ImageNet stats in BCHW format, i.e. = x * std + mean
+    for i in range(3):
+        x[:, i] = x[:, i] * std[i] + mean[i]
+    return x
+
+
+def augment_hsv(im, hgain=0.5, sgain=0.5, vgain=0.5):
+    # HSV color-space augmentation
+    if hgain or sgain or vgain:
+        r = np.random.uniform(-1, 1, 3) * [hgain, sgain, vgain] + 1  # random gains
+        hue, sat, val = cv2.split(cv2.cvtColor(im, cv2.COLOR_BGR2HSV))
+        dtype = im.dtype  # uint8
+
+        x = np.arange(0, 256, dtype=r.dtype)
+        lut_hue = ((x * r[0]) % 180).astype(dtype)
+        lut_sat = np.clip(x * r[1], 0, 255).astype(dtype)
+        lut_val = np.clip(x * r[2], 0, 255).astype(dtype)
+
+        im_hsv = cv2.merge((cv2.LUT(hue, lut_hue), cv2.LUT(sat, lut_sat), cv2.LUT(val, lut_val)))
+        cv2.cvtColor(im_hsv, cv2.COLOR_HSV2BGR, dst=im)  # no return needed
+
+
+def hist_equalize(im, clahe=True, bgr=False):
+    # Equalize histogram on BGR image 'im' with im.shape(n,m,3) and range 0-255
+    yuv = cv2.cvtColor(im, cv2.COLOR_BGR2YUV if bgr else cv2.COLOR_RGB2YUV)
+    if clahe:
+        c = cv2.createCLAHE(clipLimit=2.0, tileGridSize=(8, 8))
+        yuv[:, :, 0] = c.apply(yuv[:, :, 0])
+    else:
+        yuv[:, :, 0] = cv2.equalizeHist(yuv[:, :, 0])  # equalize Y channel histogram
+    return cv2.cvtColor(yuv, cv2.COLOR_YUV2BGR if bgr else cv2.COLOR_YUV2RGB)  # convert YUV image to RGB
+
+
+def replicate(im, labels):
+    # Replicate labels
+    h, w = im.shape[:2]
+    boxes = labels[:, 1:].astype(int)
+    x1, y1, x2, y2 = boxes.T
+    s = ((x2 - x1) + (y2 - y1)) / 2  # side length (pixels)
+    for i in s.argsort()[:round(s.size * 0.5)]:  # smallest indices
+        x1b, y1b, x2b, y2b = boxes[i]
+        bh, bw = y2b - y1b, x2b - x1b
+        yc, xc = int(random.uniform(0, h - bh)), int(random.uniform(0, w - bw))  # offset x, y
+        x1a, y1a, x2a, y2a = [xc, yc, xc + bw, yc + bh]
+        im[y1a:y2a, x1a:x2a] = im[y1b:y2b, x1b:x2b]  # im4[ymin:ymax, xmin:xmax]
+        labels = np.append(labels, [[labels[i, 0], x1a, y1a, x2a, y2a]], axis=0)
+
+    return im, labels
+
+
+def letterbox(im, new_shape=(640, 640), color=(114, 114, 114), auto=True, scaleFill=False, scaleup=True, stride=32):
+    # Resize and pad image while meeting stride-multiple constraints
+    shape = im.shape[:2]  # current shape [height, width]
+    if isinstance(new_shape, int):
+        new_shape = (new_shape, new_shape)
+
+    # Scale ratio (new / old)
+    r = min(new_shape[0] / shape[0], new_shape[1] / shape[1])
+    if not scaleup:  # only scale down, do not scale up (for better val mAP)
+        r = min(r, 1.0)
+
+    # Compute padding
+    ratio = r, r  # width, height ratios
+    new_unpad = int(round(shape[1] * r)), int(round(shape[0] * r))
+    dw, dh = new_shape[1] - new_unpad[0], new_shape[0] - new_unpad[1]  # wh padding
+    if auto:  # minimum rectangle
+        dw, dh = np.mod(dw, stride), np.mod(dh, stride)  # wh padding
+    elif scaleFill:  # stretch
+        dw, dh = 0.0, 0.0
+        new_unpad = (new_shape[1], new_shape[0])
+        ratio = new_shape[1] / shape[1], new_shape[0] / shape[0]  # width, height ratios
+
+    dw /= 2  # divide padding into 2 sides
+    dh /= 2
+
+    if shape[::-1] != new_unpad:  # resize
+        im = cv2.resize(im, new_unpad, interpolation=cv2.INTER_LINEAR)
+    top, bottom = int(round(dh - 0.1)), int(round(dh + 0.1))
+    left, right = int(round(dw - 0.1)), int(round(dw + 0.1))
+    im = cv2.copyMakeBorder(im, top, bottom, left, right, cv2.BORDER_CONSTANT, value=color)  # add border
+    return im, ratio, (dw, dh)
+
+
+def random_perspective(im,
+                       targets=(),
+                       segments=(),
+                       degrees=10,
+                       translate=.1,
+                       scale=.1,
+                       shear=10,
+                       perspective=0.0,
+                       border=(0, 0)):
+    # torchvision.transforms.RandomAffine(degrees=(-10, 10), translate=(0.1, 0.1), scale=(0.9, 1.1), shear=(-10, 10))
+    # targets = [cls, xyxy]
+
+    height = im.shape[0] + border[0] * 2  # shape(h,w,c)
+    width = im.shape[1] + border[1] * 2
+
+    # Center
+    C = np.eye(3)
+    C[0, 2] = -im.shape[1] / 2  # x translation (pixels)
+    C[1, 2] = -im.shape[0] / 2  # y translation (pixels)
+
+    # Perspective
+    P = np.eye(3)
+    P[2, 0] = random.uniform(-perspective, perspective)  # x perspective (about y)
+    P[2, 1] = random.uniform(-perspective, perspective)  # y perspective (about x)
+
+    # Rotation and Scale
+    R = np.eye(3)
+    a = random.uniform(-degrees, degrees)
+    # a += random.choice([-180, -90, 0, 90])  # add 90deg rotations to small rotations
+    s = random.uniform(1 - scale, 1 + scale)
+    # s = 2 ** random.uniform(-scale, scale)
+    R[:2] = cv2.getRotationMatrix2D(angle=a, center=(0, 0), scale=s)
+
+    # Shear
+    S = np.eye(3)
+    S[0, 1] = math.tan(random.uniform(-shear, shear) * math.pi / 180)  # x shear (deg)
+    S[1, 0] = math.tan(random.uniform(-shear, shear) * math.pi / 180)  # y shear (deg)
+
+    # Translation
+    T = np.eye(3)
+    T[0, 2] = random.uniform(0.5 - translate, 0.5 + translate) * width  # x translation (pixels)
+    T[1, 2] = random.uniform(0.5 - translate, 0.5 + translate) * height  # y translation (pixels)
+
+    # Combined rotation matrix
+    M = T @ S @ R @ P @ C  # order of operations (right to left) is IMPORTANT
+    if (border[0] != 0) or (border[1] != 0) or (M != np.eye(3)).any():  # image changed
+        if perspective:
+            im = cv2.warpPerspective(im, M, dsize=(width, height), borderValue=(114, 114, 114))
+        else:  # affine
+            im = cv2.warpAffine(im, M[:2], dsize=(width, height), borderValue=(114, 114, 114))
+
+    # Visualize
+    # import matplotlib.pyplot as plt
+    # ax = plt.subplots(1, 2, figsize=(12, 6))[1].ravel()
+    # ax[0].imshow(im[:, :, ::-1])  # base
+    # ax[1].imshow(im2[:, :, ::-1])  # warped
+
+    # Transform label coordinates
+    n = len(targets)
+    if n:
+        use_segments = any(x.any() for x in segments)
+        new = np.zeros((n, 4))
+        if use_segments:  # warp segments
+            segments = resample_segments(segments)  # upsample
+            for i, segment in enumerate(segments):
+                xy = np.ones((len(segment), 3))
+                xy[:, :2] = segment
+                xy = xy @ M.T  # transform
+                xy = xy[:, :2] / xy[:, 2:3] if perspective else xy[:, :2]  # perspective rescale or affine
+
+                # clip
+                new[i] = segment2box(xy, width, height)
+
+        else:  # warp boxes
+            xy = np.ones((n * 4, 3))
+            xy[:, :2] = targets[:, [1, 2, 3, 4, 1, 4, 3, 2]].reshape(n * 4, 2)  # x1y1, x2y2, x1y2, x2y1
+            xy = xy @ M.T  # transform
+            xy = (xy[:, :2] / xy[:, 2:3] if perspective else xy[:, :2]).reshape(n, 8)  # perspective rescale or affine
+
+            # create new boxes
+            x = xy[:, [0, 2, 4, 6]]
+            y = xy[:, [1, 3, 5, 7]]
+            new = np.concatenate((x.min(1), y.min(1), x.max(1), y.max(1))).reshape(4, n).T
+
+            # clip
+            new[:, [0, 2]] = new[:, [0, 2]].clip(0, width)
+            new[:, [1, 3]] = new[:, [1, 3]].clip(0, height)
+
+        # filter candidates
+        i = box_candidates(box1=targets[:, 1:5].T * s, box2=new.T, area_thr=0.01 if use_segments else 0.10)
+        targets = targets[i]
+        targets[:, 1:5] = new[i]
+
+    return im, targets
+
+
+def copy_paste(im, labels, segments, p=0.5):
+    # Implement Copy-Paste augmentation https://arxiv.org/abs/2012.07177, labels as nx5 np.array(cls, xyxy)
+    n = len(segments)
+    if p and n:
+        h, w, c = im.shape  # height, width, channels
+        im_new = np.zeros(im.shape, np.uint8)
+        for j in random.sample(range(n), k=round(p * n)):
+            l, s = labels[j], segments[j]
+            box = w - l[3], l[2], w - l[1], l[4]
+            ioa = bbox_ioa(box, labels[:, 1:5])  # intersection over area
+            if (ioa < 0.30).all():  # allow 30% obscuration of existing labels
+                labels = np.concatenate((labels, [[l[0], *box]]), 0)
+                segments.append(np.concatenate((w - s[:, 0:1], s[:, 1:2]), 1))
+                cv2.drawContours(im_new, [segments[j].astype(np.int32)], -1, (1, 1, 1), cv2.FILLED)
+
+        result = cv2.flip(im, 1)  # augment segments (flip left-right)
+        i = cv2.flip(im_new, 1).astype(bool)
+        im[i] = result[i]  # cv2.imwrite('debug.jpg', im)  # debug
+
+    return im, labels, segments
+
+
+def cutout(im, labels, p=0.5):
+    # Applies image cutout augmentation https://arxiv.org/abs/1708.04552
+    if random.random() < p:
+        h, w = im.shape[:2]
+        scales = [0.5] * 1 + [0.25] * 2 + [0.125] * 4 + [0.0625] * 8 + [0.03125] * 16  # image size fraction
+        for s in scales:
+            mask_h = random.randint(1, int(h * s))  # create random masks
+            mask_w = random.randint(1, int(w * s))
+
+            # box
+            xmin = max(0, random.randint(0, w) - mask_w // 2)
+            ymin = max(0, random.randint(0, h) - mask_h // 2)
+            xmax = min(w, xmin + mask_w)
+            ymax = min(h, ymin + mask_h)
+
+            # apply random color mask
+            im[ymin:ymax, xmin:xmax] = [random.randint(64, 191) for _ in range(3)]
+
+            # return unobscured labels
+            if len(labels) and s > 0.03:
+                box = np.array([xmin, ymin, xmax, ymax], dtype=np.float32)
+                ioa = bbox_ioa(box, xywhn2xyxy(labels[:, 1:5], w, h))  # intersection over area
+                labels = labels[ioa < 0.60]  # remove >60% obscured labels
+
+    return labels
+
+
+def mixup(im, labels, im2, labels2):
+    # Applies MixUp augmentation https://arxiv.org/pdf/1710.09412.pdf
+    r = np.random.beta(32.0, 32.0)  # mixup ratio, alpha=beta=32.0
+    im = (im * r + im2 * (1 - r)).astype(np.uint8)
+    labels = np.concatenate((labels, labels2), 0)
+    return im, labels
+
+
+def box_candidates(box1, box2, wh_thr=2, ar_thr=100, area_thr=0.1, eps=1e-16):  # box1(4,n), box2(4,n)
+    # Compute candidate boxes: box1 before augment, box2 after augment, wh_thr (pixels), aspect_ratio_thr, area_ratio
+    w1, h1 = box1[2] - box1[0], box1[3] - box1[1]
+    w2, h2 = box2[2] - box2[0], box2[3] - box2[1]
+    ar = np.maximum(w2 / (h2 + eps), h2 / (w2 + eps))  # aspect ratio
+    return (w2 > wh_thr) & (h2 > wh_thr) & (w2 * h2 / (w1 * h1 + eps) > area_thr) & (ar < ar_thr)  # candidates
+
+
+def classify_albumentations(
+        augment=True,
+        size=224,
+        scale=(0.08, 1.0),
+        ratio=(0.75, 1.0 / 0.75),  # 0.75, 1.33
+        hflip=0.5,
+        vflip=0.0,
+        jitter=0.4,
+        mean=IMAGENET_MEAN,
+        std=IMAGENET_STD,
+        auto_aug=False):
+    # YOLOv5 classification Albumentations (optional, only used if package is installed)
+    prefix = colorstr('albumentations: ')
+    try:
+        import albumentations as A
+        from albumentations.pytorch import ToTensorV2
+        check_version(A.__version__, '1.0.3', hard=True)  # version requirement
+        if augment:  # Resize and crop
+            T = [A.RandomResizedCrop(height=size, width=size, scale=scale, ratio=ratio)]
+            if auto_aug:
+                # TODO: implement AugMix, AutoAug & RandAug in albumentation
+                LOGGER.info(f'{prefix}auto augmentations are currently not supported')
+            else:
+                if hflip > 0:
+                    T += [A.HorizontalFlip(p=hflip)]
+                if vflip > 0:
+                    T += [A.VerticalFlip(p=vflip)]
+                if jitter > 0:
+                    color_jitter = (float(jitter),) * 3  # repeat value for brightness, contrast, satuaration, 0 hue
+                    T += [A.ColorJitter(*color_jitter, 0)]
+        else:  # Use fixed crop for eval set (reproducibility)
+            T = [A.SmallestMaxSize(max_size=size), A.CenterCrop(height=size, width=size)]
+        T += [A.Normalize(mean=mean, std=std), ToTensorV2()]  # Normalize and convert to Tensor
+        LOGGER.info(prefix + ', '.join(f'{x}'.replace('always_apply=False, ', '') for x in T if x.p))
+        return A.Compose(T)
+
+    except ImportError:  # package not installed, skip
+        LOGGER.warning(f'{prefix}⚠️ not found, install with `pip install albumentations` (recommended)')
+    except Exception as e:
+        LOGGER.info(f'{prefix}{e}')
+
+
+def classify_transforms(size=224):
+    # Transforms to apply if albumentations not installed
+    assert isinstance(size, int), f'ERROR: classify_transforms size {size} must be integer, not (list, tuple)'
+    # T.Compose([T.ToTensor(), T.Resize(size), T.CenterCrop(size), T.Normalize(IMAGENET_MEAN, IMAGENET_STD)])
+    return T.Compose([CenterCrop(size), ToTensor(), T.Normalize(IMAGENET_MEAN, IMAGENET_STD)])
+
+
+class LetterBox:
+    # YOLOv5 LetterBox class for image preprocessing, i.e. T.Compose([LetterBox(size), ToTensor()])
+    def __init__(self, size=(640, 640), auto=False, stride=32):
+        super().__init__()
+        self.h, self.w = (size, size) if isinstance(size, int) else size
+        self.auto = auto  # pass max size integer, automatically solve for short side using stride
+        self.stride = stride  # used with auto
+
+    def __call__(self, im):  # im = np.array HWC
+        imh, imw = im.shape[:2]
+        r = min(self.h / imh, self.w / imw)  # ratio of new/old
+        h, w = round(imh * r), round(imw * r)  # resized image
+        hs, ws = (math.ceil(x / self.stride) * self.stride for x in (h, w)) if self.auto else self.h, self.w
+        top, left = round((hs - h) / 2 - 0.1), round((ws - w) / 2 - 0.1)
+        im_out = np.full((self.h, self.w, 3), 114, dtype=im.dtype)
+        im_out[top:top + h, left:left + w] = cv2.resize(im, (w, h), interpolation=cv2.INTER_LINEAR)
+        return im_out
+
+
+class CenterCrop:
+    # YOLOv5 CenterCrop class for image preprocessing, i.e. T.Compose([CenterCrop(size), ToTensor()])
+    def __init__(self, size=640):
+        super().__init__()
+        self.h, self.w = (size, size) if isinstance(size, int) else size
+
+    def __call__(self, im):  # im = np.array HWC
+        imh, imw = im.shape[:2]
+        m = min(imh, imw)  # min dimension
+        top, left = (imh - m) // 2, (imw - m) // 2
+        return cv2.resize(im[top:top + m, left:left + m], (self.w, self.h), interpolation=cv2.INTER_LINEAR)
+
+
+class ToTensor:
+    # YOLOv5 ToTensor class for image preprocessing, i.e. T.Compose([LetterBox(size), ToTensor()])
+    def __init__(self, half=False):
+        super().__init__()
+        self.half = half
+
+    def __call__(self, im):  # im = np.array HWC in BGR order
+        im = np.ascontiguousarray(im.transpose((2, 0, 1))[::-1])  # HWC to CHW -> BGR to RGB -> contiguous
+        im = torch.from_numpy(im)  # to torch
+        im = im.half() if self.half else im.float()  # uint8 to fp16/32
+        im /= 255.0  # 0-255 to 0.0-1.0
+        return im

utils/autoanchor.py

+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+"""
+AutoAnchor utils
+"""
+
+import random
+
+import numpy as np
+import torch
+import yaml
+from tqdm import tqdm
+
+from utils import TryExcept
+from utils.general import LOGGER, TQDM_BAR_FORMAT, colorstr
+
+PREFIX = colorstr('AutoAnchor: ')
+
+
+def check_anchor_order(m):
+    # Check anchor order against stride order for YOLOv5 Detect() module m, and correct if necessary
+    a = m.anchors.prod(-1).mean(-1).view(-1)  # mean anchor area per output layer
+    da = a[-1] - a[0]  # delta a
+    ds = m.stride[-1] - m.stride[0]  # delta s
+    if da and (da.sign() != ds.sign()):  # same order
+        LOGGER.info(f'{PREFIX}Reversing anchor order')
+        m.anchors[:] = m.anchors.flip(0)
+
+
+@TryExcept(f'{PREFIX}ERROR')
+def check_anchors(dataset, model, thr=4.0, imgsz=640):
+    # Check anchor fit to data, recompute if necessary
+    m = model.module.model[-1] if hasattr(model, 'module') else model.model[-1]  # Detect()
+    shapes = imgsz * dataset.shapes / dataset.shapes.max(1, keepdims=True)
+    scale = np.random.uniform(0.9, 1.1, size=(shapes.shape[0], 1))  # augment scale
+    wh = torch.tensor(np.concatenate([l[:, 3:5] * s for s, l in zip(shapes * scale, dataset.labels)])).float()  # wh
+
+    def metric(k):  # compute metric
+        r = wh[:, None] / k[None]
+        x = torch.min(r, 1 / r).min(2)[0]  # ratio metric
+        best = x.max(1)[0]  # best_x
+        aat = (x > 1 / thr).float().sum(1).mean()  # anchors above threshold
+        bpr = (best > 1 / thr).float().mean()  # best possible recall
+        return bpr, aat
+
+    stride = m.stride.to(m.anchors.device).view(-1, 1, 1)  # model strides
+    anchors = m.anchors.clone() * stride  # current anchors
+    bpr, aat = metric(anchors.cpu().view(-1, 2))
+    s = f'\n{PREFIX}{aat:.2f} anchors/target, {bpr:.3f} Best Possible Recall (BPR). '
+    if bpr > 0.98:  # threshold to recompute
+        LOGGER.info(f'{s}Current anchors are a good fit to dataset ✅')
+    else:
+        LOGGER.info(f'{s}Anchors are a poor fit to dataset ⚠️, attempting to improve...')
+        na = m.anchors.numel() // 2  # number of anchors
+        anchors = kmean_anchors(dataset, n=na, img_size=imgsz, thr=thr, gen=1000, verbose=False)
+        new_bpr = metric(anchors)[0]
+        if new_bpr > bpr:  # replace anchors
+            anchors = torch.tensor(anchors, device=m.anchors.device).type_as(m.anchors)
+            m.anchors[:] = anchors.clone().view_as(m.anchors)
+            check_anchor_order(m)  # must be in pixel-space (not grid-space)
+            m.anchors /= stride
+            s = f'{PREFIX}Done ✅ (optional: update model *.yaml to use these anchors in the future)'
+        else:
+            s = f'{PREFIX}Done ⚠️ (original anchors better than new anchors, proceeding with original anchors)'
+        LOGGER.info(s)
+
+
+def kmean_anchors(dataset='./data/coco128.yaml', n=9, img_size=640, thr=4.0, gen=1000, verbose=True):
+    """ Creates kmeans-evolved anchors from training dataset
+
+        Arguments:
+            dataset: path to data.yaml, or a loaded dataset
+            n: number of anchors
+            img_size: image size used for training
+            thr: anchor-label wh ratio threshold hyperparameter hyp['anchor_t'] used for training, default=4.0
+            gen: generations to evolve anchors using genetic algorithm
+            verbose: print all results
+
+        Return:
+            k: kmeans evolved anchors
+
+        Usage:
+            from utils.autoanchor import *; _ = kmean_anchors()
+    """
+    from scipy.cluster.vq import kmeans
+
+    npr = np.random
+    thr = 1 / thr
+
+    def metric(k, wh):  # compute metrics
+        r = wh[:, None] / k[None]
+        x = torch.min(r, 1 / r).min(2)[0]  # ratio metric
+        # x = wh_iou(wh, torch.tensor(k))  # iou metric
+        return x, x.max(1)[0]  # x, best_x
+
+    def anchor_fitness(k):  # mutation fitness
+        _, best = metric(torch.tensor(k, dtype=torch.float32), wh)
+        return (best * (best > thr).float()).mean()  # fitness
+
+    def print_results(k, verbose=True):
+        k = k[np.argsort(k.prod(1))]  # sort small to large
+        x, best = metric(k, wh0)
+        bpr, aat = (best > thr).float().mean(), (x > thr).float().mean() * n  # best possible recall, anch > thr
+        s = f'{PREFIX}thr={thr:.2f}: {bpr:.4f} best possible recall, {aat:.2f} anchors past thr\n' \
+            f'{PREFIX}n={n}, img_size={img_size}, metric_all={x.mean():.3f}/{best.mean():.3f}-mean/best, ' \
+            f'past_thr={x[x > thr].mean():.3f}-mean: '
+        for x in k:
+            s += '%i,%i, ' % (round(x[0]), round(x[1]))
+        if verbose:
+            LOGGER.info(s[:-2])
+        return k
+
+    if isinstance(dataset, str):  # *.yaml file
+        with open(dataset, errors='ignore') as f:
+            data_dict = yaml.safe_load(f)  # model dict
+        from utils.dataloaders import LoadImagesAndLabels
+        dataset = LoadImagesAndLabels(data_dict['train'], augment=True, rect=True)
+
+    # Get label wh
+    shapes = img_size * dataset.shapes / dataset.shapes.max(1, keepdims=True)
+    wh0 = np.concatenate([l[:, 3:5] * s for s, l in zip(shapes, dataset.labels)])  # wh
+
+    # Filter
+    i = (wh0 < 3.0).any(1).sum()
+    if i:
+        LOGGER.info(f'{PREFIX}WARNING ⚠️ Extremely small objects found: {i} of {len(wh0)} labels are <3 pixels in size')
+    wh = wh0[(wh0 >= 2.0).any(1)].astype(np.float32)  # filter > 2 pixels
+    # wh = wh * (npr.rand(wh.shape[0], 1) * 0.9 + 0.1)  # multiply by random scale 0-1
+
+    # Kmeans init
+    try:
+        LOGGER.info(f'{PREFIX}Running kmeans for {n} anchors on {len(wh)} points...')
+        assert n <= len(wh)  # apply overdetermined constraint
+        s = wh.std(0)  # sigmas for whitening
+        k = kmeans(wh / s, n, iter=30)[0] * s  # points
+        assert n == len(k)  # kmeans may return fewer points than requested if wh is insufficient or too similar
+    except Exception:
+        LOGGER.warning(f'{PREFIX}WARNING ⚠️ switching strategies from kmeans to random init')
+        k = np.sort(npr.rand(n * 2)).reshape(n, 2) * img_size  # random init
+    wh, wh0 = (torch.tensor(x, dtype=torch.float32) for x in (wh, wh0))
+    k = print_results(k, verbose=False)
+
+    # Plot
+    # k, d = [None] * 20, [None] * 20
+    # for i in tqdm(range(1, 21)):
+    #     k[i-1], d[i-1] = kmeans(wh / s, i)  # points, mean distance
+    # fig, ax = plt.subplots(1, 2, figsize=(14, 7), tight_layout=True)
+    # ax = ax.ravel()
+    # ax[0].plot(np.arange(1, 21), np.array(d) ** 2, marker='.')
+    # fig, ax = plt.subplots(1, 2, figsize=(14, 7))  # plot wh
+    # ax[0].hist(wh[wh[:, 0]<100, 0],400)
+    # ax[1].hist(wh[wh[:, 1]<100, 1],400)
+    # fig.savefig('wh.png', dpi=200)
+
+    # Evolve
+    f, sh, mp, s = anchor_fitness(k), k.shape, 0.9, 0.1  # fitness, generations, mutation prob, sigma
+    pbar = tqdm(range(gen), bar_format=TQDM_BAR_FORMAT)  # progress bar
+    for _ in pbar:
+        v = np.ones(sh)
+        while (v == 1).all():  # mutate until a change occurs (prevent duplicates)
+            v = ((npr.random(sh) < mp) * random.random() * npr.randn(*sh) * s + 1).clip(0.3, 3.0)
+        kg = (k.copy() * v).clip(min=2.0)
+        fg = anchor_fitness(kg)
+        if fg > f:
+            f, k = fg, kg.copy()
+            pbar.desc = f'{PREFIX}Evolving anchors with Genetic Algorithm: fitness = {f:.4f}'
+            if verbose:
+                print_results(k, verbose)
+
+    return print_results(k).astype(np.float32)

utils/autobatch.py

+# YOLOv5 🚀 by Ultralytics, GPL-3.0 license
+"""
+Auto-batch utils
+"""
+
+from copy import deepcopy
+
+import numpy as np
+import torch
+
+from utils.general import LOGGER, colorstr
+from utils.torch_utils import profile
+
+
+def check_train_batch_size(model, imgsz=640, amp=True):
+    # Check YOLOv5 training batch size
+    with torch.cuda.amp.autocast(amp):
+        return autobatch(deepcopy(model).train(), imgsz)  # compute optimal batch size
+
+
+def autobatch(model, imgsz=640, fraction=0.8, batch_size=16):
+    # Automatically estimate best YOLOv5 batch size to use `fraction` of available CUDA memory
+    # Usage:
+    #     import torch
+    #     from utils.autobatch import autobatch
+    #     model = torch.hub.load('ultralytics/yolov5', 'yolov5s', autoshape=False)
+    #     print(autobatch(model))
+
+    # Check device
+    prefix = colorstr('AutoBatch: ')
+    LOGGER.info(f'{prefix}Computing optimal batch size for --imgsz {imgsz}')
+    device = next(model.parameters()).device  # get model device
+    if device.type == 'cpu':
+        LOGGER.info(f'{prefix}CUDA not detected, using default CPU batch-size {batch_size}')
+        return batch_size
+    if torch.backends.cudnn.benchmark:
+        LOGGER.info(f'{prefix} ⚠️ Requires torch.backends.cudnn.benchmark=False, using default batch-size {batch_size}')
+        return batch_size
+
+    # Inspect CUDA memory
+    gb = 1 << 30  # bytes to GiB (1024 ** 3)
+    d = str(device).upper()  # 'CUDA:0'
+    properties = torch.cuda.get_device_properties(device)  # device properties
+    t = properties.total_memory / gb  # GiB total
+    r = torch.cuda.memory_reserved(device) / gb  # GiB reserved
+    a = torch.cuda.memory_allocated(device) / gb  # GiB allocated
+    f = t - (r + a)  # GiB free
+    LOGGER.info(f'{prefix}{d} ({properties.name}) {t:.2f}G total, {r:.2f}G reserved, {a:.2f}G allocated, {f:.2f}G free')
+
+    # Profile batch sizes
+    batch_sizes = [1, 2, 4, 8, 16]
+    try:
+        img = [torch.empty(b, 3, imgsz, imgsz) for b in batch_sizes]
+        results = profile(img, model, n=3, device=device)
+    except Exception as e:
+        LOGGER.warning(f'{prefix}{e}')
+
+    # Fit a solution
+    y = [x[2] for x in results if x]  # memory [2]
+    p = np.polyfit(batch_sizes[:len(y)], y, deg=1)  # first degree polynomial fit
+    b = int((f * fraction - p[1]) / p[0])  # y intercept (optimal batch size)
+    if None in results:  # some sizes failed
+        i = results.index(None)  # first fail index
+        if b >= batch_sizes[i]:  # y intercept above failure point
+            b = batch_sizes[max(i - 1, 0)]  # select prior safe point
+    if b < 1 or b > 1024:  # b outside of safe range
+        b = batch_size
+        LOGGER.warning(f'{prefix}WARNING ⚠️ CUDA anomaly detected, recommend restart environment and retry command.')
+
+    fraction = (np.polyval(p, b) + r + a) / t  # actual fraction predicted
+    LOGGER.info(f'{prefix}Using batch-size {b} for {d} {t * fraction:.2f}G/{t:.2f}G ({fraction * 100:.0f}%) ✅')
+    return b