Skip to content

GitLab

Commit a4372e17 authored by huchen's avatar huchen
Browse files

Merge branch 'hepj_work' into 'main' 

增加conformer代码

See merge request dcutoolkit/deeplearing/dlexamples_new!7
parents 7f99c1c3 142dcf29
Hide whitespace changes
Inline Side-by-side
Showing with 2250 additions and 0 deletions
PyTorch/NLP/Conformer-main/LICENSE 0 → 100644
View file @ a4372e17
Apache License
Version 2.0, January 2004
http://www.apache.org/licenses/
TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
1. Definitions.
"License" shall mean the terms and conditions for use, reproduction,
and distribution as defined by Sections 1 through 9 of this document.
"Licensor" shall mean the copyright owner or entity authorized by
the copyright owner that is granting the License.
"Legal Entity" shall mean the union of the acting entity and all
other entities that control, are controlled by, or are under common
control with that entity. For the purposes of this definition,
"control" means (i) the power, direct or indirect, to cause the
direction or management of such entity, whether by contract or
otherwise, or (ii) ownership of fifty percent (50%) or more of the
outstanding shares, or (iii) beneficial ownership of such entity.
"You" (or "Your") shall mean an individual or Legal Entity
exercising permissions granted by this License.
"Source" form shall mean the preferred form for making modifications,
including but not limited to software source code, documentation
source, and configuration files.
"Object" form shall mean any form resulting from mechanical
transformation or translation of a Source form, including but
not limited to compiled object code, generated documentation,
and conversions to other media types.
"Work" shall mean the work of authorship, whether in Source or
Object form, made available under the License, as indicated by a
copyright notice that is included in or attached to the work
(an example is provided in the Appendix below).
"Derivative Works" shall mean any work, whether in Source or Object
form, that is based on (or derived from) the Work and for which the
editorial revisions, annotations, elaborations, or other modifications
represent, as a whole, an original work of authorship. For the purposes
of this License, Derivative Works shall not include works that remain
separable from, or merely link (or bind by name) to the interfaces of,
the Work and Derivative Works thereof.
"Contribution" shall mean any work of authorship, including
the original version of the Work and any modifications or additions
to that Work or Derivative Works thereof, that is intentionally
submitted to Licensor for inclusion in the Work by the copyright owner
or by an individual or Legal Entity authorized to submit on behalf of
the copyright owner. For the purposes of this definition, "submitted"
means any form of electronic, verbal, or written communication sent
to the Licensor or its representatives, including but not limited to
communication on electronic mailing lists, source code control systems,
and issue tracking systems that are managed by, or on behalf of, the
Licensor for the purpose of discussing and improving the Work, but
excluding communication that is conspicuously marked or otherwise
designated in writing by the copyright owner as "Not a Contribution."
"Contributor" shall mean Licensor and any individual or Legal Entity
on behalf of whom a Contribution has been received by Licensor and
subsequently incorporated within the Work.
2. Grant of Copyright License. Subject to the terms and conditions of
this License, each Contributor hereby grants to You a perpetual,
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
copyright license to reproduce, prepare Derivative Works of,
publicly display, publicly perform, sublicense, and distribute the
Work and such Derivative Works in Source or Object form.
3. Grant of Patent License. Subject to the terms and conditions of
this License, each Contributor hereby grants to You a perpetual,
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
(except as stated in this section) patent license to make, have made,
use, offer to sell, sell, import, and otherwise transfer the Work,
where such license applies only to those patent claims licensable
by such Contributor that are necessarily infringed by their
Contribution(s) alone or by combination of their Contribution(s)
with the Work to which such Contribution(s) was submitted. If You
institute patent litigation against any entity (including a
cross-claim or counterclaim in a lawsuit) alleging that the Work
or a Contribution incorporated within the Work constitutes direct
or contributory patent infringement, then any patent licenses
granted to You under this License for that Work shall terminate
as of the date such litigation is filed.
4. Redistribution. You may reproduce and distribute copies of the
Work or Derivative Works thereof in any medium, with or without
modifications, and in Source or Object form, provided that You
meet the following conditions:
(a) You must give any other recipients of the Work or
Derivative Works a copy of this License; and
(b) You must cause any modified files to carry prominent notices
stating that You changed the files; and
(c) You must retain, in the Source form of any Derivative Works
that You distribute, all copyright, patent, trademark, and
attribution notices from the Source form of the Work,
excluding those notices that do not pertain to any part of
the Derivative Works; and
(d) If the Work includes a "NOTICE" text file as part of its
distribution, then any Derivative Works that You distribute must
include a readable copy of the attribution notices contained
within such NOTICE file, excluding those notices that do not
pertain to any part of the Derivative Works, in at least one
of the following places: within a NOTICE text file distributed
as part of the Derivative Works; within the Source form or
documentation, if provided along with the Derivative Works; or,
within a display generated by the Derivative Works, if and
wherever such third-party notices normally appear. The contents
of the NOTICE file are for informational purposes only and
do not modify the License. You may add Your own attribution
notices within Derivative Works that You distribute, alongside
or as an addendum to the NOTICE text from the Work, provided
that such additional attribution notices cannot be construed
as modifying the License.
You may add Your own copyright statement to Your modifications and
may provide additional or different license terms and conditions
for use, reproduction, or distribution of Your modifications, or
for any such Derivative Works as a whole, provided Your use,
reproduction, and distribution of the Work otherwise complies with
the conditions stated in this License.
5. Submission of Contributions. Unless You explicitly state otherwise,
any Contribution intentionally submitted for inclusion in the Work
by You to the Licensor shall be under the terms and conditions of
this License, without any additional terms or conditions.
Notwithstanding the above, nothing herein shall supersede or modify
the terms of any separate license agreement you may have executed
with Licensor regarding such Contributions.
6. Trademarks. This License does not grant permission to use the trade
names, trademarks, service marks, or product names of the Licensor,
except as required for reasonable and customary use in describing the
origin of the Work and reproducing the content of the NOTICE file.
7. Disclaimer of Warranty. Unless required by applicable law or
agreed to in writing, Licensor provides the Work (and each
Contributor provides its Contributions) on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
implied, including, without limitation, any warranties or conditions
of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
PARTICULAR PURPOSE. You are solely responsible for determining the
appropriateness of using or redistributing the Work and assume any
risks associated with Your exercise of permissions under this License.
8. Limitation of Liability. In no event and under no legal theory,
whether in tort (including negligence), contract, or otherwise,
unless required by applicable law (such as deliberate and grossly
negligent acts) or agreed to in writing, shall any Contributor be
liable to You for damages, including any direct, indirect, special,
incidental, or consequential damages of any character arising as a
result of this License or out of the use or inability to use the
Work (including but not limited to damages for loss of goodwill,
work stoppage, computer failure or malfunction, or any and all
other commercial damages or losses), even if such Contributor
has been advised of the possibility of such damages.
9. Accepting Warranty or Additional Liability. While redistributing
the Work or Derivative Works thereof, You may choose to offer,
and charge a fee for, acceptance of support, warranty, indemnity,
or other liability obligations and/or rights consistent with this
License. However, in accepting such obligations, You may act only
on Your own behalf and on Your sole responsibility, not on behalf
of any other Contributor, and only if You agree to indemnify,
defend, and hold each Contributor harmless for any liability
incurred by, or claims asserted against, such Contributor by reason
of your accepting any such warranty or additional liability.
END OF TERMS AND CONDITIONS
APPENDIX: How to apply the Apache License to your work.
To apply the Apache License to your work, attach the following
boilerplate notice, with the fields enclosed by brackets "[]"
replaced with your own identifying information. (Don't include
the brackets!) The text should be enclosed in the appropriate
comment syntax for the file format. We also recommend that a
file or class name and description of purpose be included on the
same "printed page" as the copyright notice for easier
identification within third-party archives.
Copyright [yyyy] [name of copyright owner]
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
PyTorch/NLP/Conformer-main/README.md 0 → 100644
View file @ a4372e17
## 简介
使用pytorch框架计算conformer网络
一些一阶查看README_ORIGIN.md
[conformer的GIT网址](https://github.com/pengzhiliang/Conformer)
## 运行前准备
```
#修改_amp_state.py：(python3.6/site-packages/apex/amp/_amp_state.py)
if TORCH_MAJOR == 1 and TORCH_MINOR < 8:
from torch._six import container_abcs
else:
import collections.abc as container_abcs
改为：
if TORCH_MAJOR == 1 and TORCH_MINOR < 8:
#from torch._six import container_abcs
import collections.abc as container_abcs
else:
import collections.abc as container_abcs
```
```
#修改helpers.py ：（python3.6/site-packages/timm/models/layers/helpers.py）
修改：
from torch._six import container_abcs
改为：
import collections.abc as container_abcs
```
## 数据集地址
昆山服务器存有数据集，地址：
/public/software/apps/DeepLearning/Data/ImageNet-pytorch
### 单卡
```
#启动
./run1.sh
```
sh脚本中--nnodes 为机器数 ，--nproc_per_node每个机器显卡数目，
对于python参数：
--num_workers 为显卡数，--data-path为数据路径，--output_dir为输出文件夹
## 多卡
```
#运行
./run4.sh
```
PyTorch/NLP/Conformer-main/README_ORIGIN.md 0 → 100644
View file @ a4372e17
# Conformer: Local Features Coupling Global Representations for Visual Recognition
**Accpeted to ICCV21!**
This repository is built upon [DeiT](https://github.com/facebookresearch/deit), [timm](https://github.com/rwightman/pytorch-image-models), and [mmdetction](https://github.com/open-mmlab/mmdetection).
# Introduction
Within Convolutional Neural Network (CNN), the convolution operations are good at extracting local features but experience difficulty to capture global representations.
Within visual transformer, the cascaded self-attention modules can capture long-distance feature dependencies but unfortunately deteriorate local feature details.
In this paper, we propose a hybrid network structure, termed Conformer, to take advantage of convolutional operations and self-attention mechanisms for enhanced representation learning.
Conformer roots in the Feature Coupling Unit (FCU), which fuses local features and global representations under different resolutions in an interactive fashion.
Conformer adopts a concurrent structure so that local features and global representations are retained to the maximum extent.
Experiments show that Conformer, under the comparable parameter complexity, outperforms the visual transformer (DeiT-B) by 2.3\% on ImageNet.
On MSCOCO, it outperforms ResNet-101 by 3.7\% and 3.6\% mAPs for object detection and instance segmentation, respectively, demonstrating the great potential to be a general backbone network.
The basic architecture of the Conformer is shown as following:
![](figures/network.png)
We also show the comparison of feature maps of CNN (ResNet-101), Visual Transformer (DeiT-S), and the proposed Conformer as following.
The patch embeddings in transformer are reshaped to feature maps for visualization. While CNN activates discriminative local regions ($e.g.$, the peacock's head in (a) and tail in (e)),
the CNN branch of Conformer takes advantage of global cues from the visual transformer and thereby activates complete object ($e.g.$, full extent of the peacock in (b) and (f)).
Compared with CNN, local feature details of the visual transformer are deteriorated ($e.g.$, (c) and (g)). In contrast,
the transformer branch of Conformer retains the local feature details from CNN while depressing the background ($e.g.$,
the peacock contours in (d) and (h) are more complete than those in(c) and (g).
![](figures/feature_maps.png)
# Getting started
## Install
First, install PyTorch 1.7.0+ and torchvision 0.8.1+ and [pytorch-image-models 0.3.2](https://github.com/rwightman/pytorch-image-models):
```
conda install -c pytorch pytorch torchvision
pip install timm==0.3.2
```
## Data preparation
Download and extract ImageNet train and val images from http://image-net.org/.
The directory structure is the standard layout for the torchvision [`datasets.ImageFolder`](https://pytorch.org/docs/stable/torchvision/datasets.html#imagefolder), and the training and validation data is expected to be in the `train/` folder and `val` folder respectively:
```
/path/to/imagenet/
train/
class1/
img1.jpeg
class2/
img2.jpeg
val/
class1/
img3.jpeg
class/2
img4.jpeg
```
## Training and test
### Training
To train Conformer-S on ImageNet on a single node with 8 gpus for 300 epochs run:
```
export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7
OUTPUT='./output/Conformer_small_patch16_batch_1024_lr1e-3_300epochs'
python -m torch.distributed.launch --master_port 50130 --nproc_per_node=8 --use_env main.py \
--model Conformer_small_patch16 \
--data-set IMNET \
--batch-size 128 \
--lr 0.001 \
--num_workers 4 \
--data-path /data/user/Dataset/ImageNet_ILSVRC2012/ \
--output_dir ${OUTPUT} \
--epochs 300
```
### Test
To test Conformer-S on ImageNet on a single gpu run:
```
CUDA_VISIBLE_DEVICES=0, python main.py --model Conformer_small_patch16 --eval --batch-size 64 \
--input-size 224 \
--data-set IMNET \
--num_workers 4 \
--data-path /data/user/Dataset/ImageNet_ILSVRC2012/ \
--epochs 100 \
--resume ../Conformer_small_patch16.pth
```
# Model zoo
| Model | Parameters | MACs | Top-1 Acc | Link |
| ------------ | ---------- | ------ | --------- | ---- |
| Conformer-Ti | 23.5 M | 5.2 G | 81.3 % | [baidu](https://pan.baidu.com/s/12AblBmhUu5gnYsPjnDE_Jg)(code: hzhm) [google](https://drive.google.com/file/d/19SxGhKcWOR5oQSxNUWUM2MGYiaWMrF1z/view?usp=sharing) |
| Conformer-S | 37.7 M | 10.6 G | 83.4 % | [baidu](https://pan.baidu.com/s/1kYOZ9mRP5fvujH6snsOjew)(code: qvu8) [google](https://drive.google.com/file/d/1mpOlbLaVxOfEwV4-ha78j_1Ebqzj2B83/view?usp=sharing) |
| Conformer-B | 83.3 M | 23.3 G | 84.1 % | [baidu](https://pan.baidu.com/s/1FL5XDAqHoimpUxNSunKq0w)(code: b4z9) [google](https://drive.google.com/file/d/1oeQ9LSOGKEUaYGu7WTlUGl3KDsQIi0MA/view?usp=sharing) |
# Citation
```
@article{peng2021conformer,
title={Conformer: Local Features Coupling Global Representations for Visual Recognition},
author={Zhiliang Peng and Wei Huang and Shanzhi Gu and Lingxi Xie and Yaowei Wang and Jianbin Jiao and Qixiang Ye},
journal={arXiv preprint arXiv:2105.03889},
year={2021},
}
```
PyTorch/NLP/Conformer-main/conformer.py 0 → 100644
View file @ a4372e17
import torch
import torch.nn as nn
import torch.nn.functional as F
from functools import partial
from timm.models.layers import DropPath, trunc_normal_
class Mlp(nn.Module):
def __init__(self, in_features, hidden_features=None, out_features=None, act_layer=nn.GELU, drop=0.):
super().__init__()
out_features = out_features or in_features
hidden_features = hidden_features or in_features
self.fc1 = nn.Linear(in_features, hidden_features)
self.act = act_layer()
self.fc2 = nn.Linear(hidden_features, out_features)
self.drop = nn.Dropout(drop)
def forward(self, x):
x = self.fc1(x)
x = self.act(x)
x = self.drop(x)
x = self.fc2(x)
x = self.drop(x)
return x
class Attention(nn.Module):
def __init__(self, dim, num_heads=8, qkv_bias=False, qk_scale=None, attn_drop=0., proj_drop=0.):
super().__init__()
self.num_heads = num_heads
head_dim = dim // num_heads
# NOTE scale factor was wrong in my original version, can set manually to be compat with prev weights
self.scale = qk_scale or head_dim ** -0.5
self.qkv = nn.Linear(dim, dim * 3, bias=qkv_bias)
self.attn_drop = nn.Dropout(attn_drop)
self.proj = nn.Linear(dim, dim)
self.proj_drop = nn.Dropout(proj_drop)
def forward(self, x):
B, N, C = x.shape
qkv = self.qkv(x).reshape(B, N, 3, self.num_heads, C // self.num_heads).permute(2, 0, 3, 1, 4)
q, k, v = qkv[0], qkv[1], qkv[2] # make torchscript happy (cannot use tensor as tuple)
attn = (q @ k.transpose(-2, -1)) * self.scale
attn = attn.softmax(dim=-1)
attn = self.attn_drop(attn)
x = (attn @ v).transpose(1, 2).reshape(B, N, C)
x = self.proj(x)
x = self.proj_drop(x)
return x
class Block(nn.Module):
def __init__(self, dim, num_heads, mlp_ratio=4., qkv_bias=False, qk_scale=None, drop=0., attn_drop=0.,
drop_path=0., act_layer=nn.GELU, norm_layer=partial(nn.LayerNorm, eps=1e-6)):
super().__init__()
self.norm1 = norm_layer(dim)
self.attn = Attention(
dim, num_heads=num_heads, qkv_bias=qkv_bias, qk_scale=qk_scale, attn_drop=attn_drop, proj_drop=drop)
# NOTE: drop path for stochastic depth, we shall see if this is better than dropout here
self.drop_path = DropPath(drop_path) if drop_path > 0. else nn.Identity()
self.norm2 = norm_layer(dim)
mlp_hidden_dim = int(dim * mlp_ratio)
self.mlp = Mlp(in_features=dim, hidden_features=mlp_hidden_dim, act_layer=act_layer, drop=drop)
def forward(self, x):
x = x + self.drop_path(self.attn(self.norm1(x)))
x = x + self.drop_path(self.mlp(self.norm2(x)))
return x
class ConvBlock(nn.Module):
def __init__(self, inplanes, outplanes, stride=1, res_conv=False, act_layer=nn.ReLU, groups=1,
norm_layer=partial(nn.BatchNorm2d, eps=1e-6), drop_block=None, drop_path=None):
super(ConvBlock, self).__init__()
expansion = 4
med_planes = outplanes // expansion
self.conv1 = nn.Conv2d(inplanes, med_planes, kernel_size=1, stride=1, padding=0, bias=False)
self.bn1 = norm_layer(med_planes)
self.act1 = act_layer(inplace=True)
self.conv2 = nn.Conv2d(med_planes, med_planes, kernel_size=3, stride=stride, groups=groups, padding=1, bias=False)
self.bn2 = norm_layer(med_planes)
self.act2 = act_layer(inplace=True)
self.conv3 = nn.Conv2d(med_planes, outplanes, kernel_size=1, stride=1, padding=0, bias=False)
self.bn3 = norm_layer(outplanes)
self.act3 = act_layer(inplace=True)
if res_conv:
self.residual_conv = nn.Conv2d(inplanes, outplanes, kernel_size=1, stride=stride, padding=0, bias=False)
self.residual_bn = norm_layer(outplanes)
self.res_conv = res_conv
self.drop_block = drop_block
self.drop_path = drop_path
def zero_init_last_bn(self):
nn.init.zeros_(self.bn3.weight)
def forward(self, x, x_t=None, return_x_2=True):
residual = x
x = self.conv1(x)
x = self.bn1(x)
if self.drop_block is not None:
x = self.drop_block(x)
x = self.act1(x)
x = self.conv2(x) if x_t is None else self.conv2(x + x_t)
x = self.bn2(x)
if self.drop_block is not None:
x = self.drop_block(x)
x2 = self.act2(x)
x = self.conv3(x2)
x = self.bn3(x)
if self.drop_block is not None:
x = self.drop_block(x)
if self.drop_path is not None:
x = self.drop_path(x)
if self.res_conv:
residual = self.residual_conv(residual)
residual = self.residual_bn(residual)
x += residual
x = self.act3(x)
if return_x_2:
return x, x2
else:
return x
class FCUDown(nn.Module):
""" CNN feature maps -> Transformer patch embeddings
"""
def __init__(self, inplanes, outplanes, dw_stride, act_layer=nn.GELU,
norm_layer=partial(nn.LayerNorm, eps=1e-6)):
super(FCUDown, self).__init__()
self.dw_stride = dw_stride
self.conv_project = nn.Conv2d(inplanes, outplanes, kernel_size=1, stride=1, padding=0)
self.sample_pooling = nn.AvgPool2d(kernel_size=dw_stride, stride=dw_stride)
self.ln = norm_layer(outplanes)
self.act = act_layer()
def forward(self, x, x_t):
x = self.conv_project(x) # [N, C, H, W]
x = self.sample_pooling(x).flatten(2).transpose(1, 2)
x = self.ln(x)
x = self.act(x)
x = torch.cat([x_t[:, 0][:, None, :], x], dim=1)
return x
class FCUUp(nn.Module):
""" Transformer patch embeddings -> CNN feature maps
"""
def __init__(self, inplanes, outplanes, up_stride, act_layer=nn.ReLU,
norm_layer=partial(nn.BatchNorm2d, eps=1e-6),):
super(FCUUp, self).__init__()
self.up_stride = up_stride
self.conv_project = nn.Conv2d(inplanes, outplanes, kernel_size=1, stride=1, padding=0)
self.bn = norm_layer(outplanes)
self.act = act_layer()
def forward(self, x, H, W):
B, _, C = x.shape
# [N, 197, 384] -> [N, 196, 384] -> [N, 384, 196] -> [N, 384, 14, 14]
x_r = x[:, 1:].transpose(1, 2).reshape(B, C, H, W)
x_r = self.act(self.bn(self.conv_project(x_r)))
return F.interpolate(x_r, size=(H * self.up_stride, W * self.up_stride))
class Med_ConvBlock(nn.Module):
""" special case for Convblock with down sampling,
"""
def __init__(self, inplanes, act_layer=nn.ReLU, groups=1, norm_layer=partial(nn.BatchNorm2d, eps=1e-6),
drop_block=None, drop_path=None):
super(Med_ConvBlock, self).__init__()
expansion = 4
med_planes = inplanes // expansion
self.conv1 = nn.Conv2d(inplanes, med_planes, kernel_size=1, stride=1, padding=0, bias=False)
self.bn1 = norm_layer(med_planes)
self.act1 = act_layer(inplace=True)
self.conv2 = nn.Conv2d(med_planes, med_planes, kernel_size=3, stride=1, groups=groups, padding=1, bias=False)
self.bn2 = norm_layer(med_planes)
self.act2 = act_layer(inplace=True)
self.conv3 = nn.Conv2d(med_planes, inplanes, kernel_size=1, stride=1, padding=0, bias=False)
self.bn3 = norm_layer(inplanes)
self.act3 = act_layer(inplace=True)
self.drop_block = drop_block
self.drop_path = drop_path
def zero_init_last_bn(self):
nn.init.zeros_(self.bn3.weight)
def forward(self, x):
residual = x
x = self.conv1(x)
x = self.bn1(x)
if self.drop_block is not None:
x = self.drop_block(x)
x = self.act1(x)
x = self.conv2(x)
x = self.bn2(x)
if self.drop_block is not None:
x = self.drop_block(x)
x = self.act2(x)
x = self.conv3(x)
x = self.bn3(x)
if self.drop_block is not None:
x = self.drop_block(x)
if self.drop_path is not None:
x = self.drop_path(x)
x += residual
x = self.act3(x)
return x
class ConvTransBlock(nn.Module):
"""
Basic module for ConvTransformer, keep feature maps for CNN block and patch embeddings for transformer encoder block
"""
def __init__(self, inplanes, outplanes, res_conv, stride, dw_stride, embed_dim, num_heads=12, mlp_ratio=4.,
qkv_bias=False, qk_scale=None, drop_rate=0., attn_drop_rate=0., drop_path_rate=0.,
last_fusion=False, num_med_block=0, groups=1):
super(ConvTransBlock, self).__init__()
expansion = 4
self.cnn_block = ConvBlock(inplanes=inplanes, outplanes=outplanes, res_conv=res_conv, stride=stride, groups=groups)
if last_fusion:
self.fusion_block = ConvBlock(inplanes=outplanes, outplanes=outplanes, stride=2, res_conv=True, groups=groups)
else:
self.fusion_block = ConvBlock(inplanes=outplanes, outplanes=outplanes, groups=groups)
if num_med_block > 0:
self.med_block = []
for i in range(num_med_block):
self.med_block.append(Med_ConvBlock(inplanes=outplanes, groups=groups))
self.med_block = nn.ModuleList(self.med_block)
self.squeeze_block = FCUDown(inplanes=outplanes // expansion, outplanes=embed_dim, dw_stride=dw_stride)
self.expand_block = FCUUp(inplanes=embed_dim, outplanes=outplanes // expansion, up_stride=dw_stride)
self.trans_block = Block(
dim=embed_dim, num_heads=num_heads, mlp_ratio=mlp_ratio, qkv_bias=qkv_bias, qk_scale=qk_scale,
drop=drop_rate, attn_drop=attn_drop_rate, drop_path=drop_path_rate)
self.dw_stride = dw_stride
self.embed_dim = embed_dim
self.num_med_block = num_med_block
self.last_fusion = last_fusion
def forward(self, x, x_t):
x, x2 = self.cnn_block(x)
_, _, H, W = x2.shape
x_st = self.squeeze_block(x2, x_t)
x_t = self.trans_block(x_st + x_t)
if self.num_med_block > 0:
for m in self.med_block:
x = m(x)
x_t_r = self.expand_block(x_t, H // self.dw_stride, W // self.dw_stride)
x = self.fusion_block(x, x_t_r, return_x_2=False)
return x, x_t
class Conformer(nn.Module):
def __init__(self, patch_size=16, in_chans=3, num_classes=1000, base_channel=64, channel_ratio=4, num_med_block=0,
embed_dim=768, depth=12, num_heads=12, mlp_ratio=4., qkv_bias=False, qk_scale=None,
drop_rate=0., attn_drop_rate=0., drop_path_rate=0.):
# Transformer
super().__init__()
self.num_classes = num_classes
self.num_features = self.embed_dim = embed_dim # num_features for consistency with other models
assert depth % 3 == 0
self.cls_token = nn.Parameter(torch.zeros(1, 1, embed_dim))
self.trans_dpr = [x.item() for x in torch.linspace(0, drop_path_rate, depth)] # stochastic depth decay rule
# Classifier head
self.trans_norm = nn.LayerNorm(embed_dim)
self.trans_cls_head = nn.Linear(embed_dim, num_classes) if num_classes > 0 else nn.Identity()
self.pooling = nn.AdaptiveAvgPool2d(1)
self.conv_cls_head = nn.Linear(int(256 * channel_ratio), num_classes)
# Stem stage: get the feature maps by conv block (copied form resnet.py)
self.conv1 = nn.Conv2d(in_chans, 64, kernel_size=7, stride=2, padding=3, bias=False) # 1 / 2 [112, 112]
self.bn1 = nn.BatchNorm2d(64)
self.act1 = nn.ReLU(inplace=True)
self.maxpool = nn.MaxPool2d(kernel_size=3, stride=2, padding=1) # 1 / 4 [56, 56]
# 1 stage
stage_1_channel = int(base_channel * channel_ratio)
trans_dw_stride = patch_size // 4
self.conv_1 = ConvBlock(inplanes=64, outplanes=stage_1_channel, res_conv=True, stride=1)
self.trans_patch_conv = nn.Conv2d(64, embed_dim, kernel_size=trans_dw_stride, stride=trans_dw_stride, padding=0)
self.trans_1 = Block(dim=embed_dim, num_heads=num_heads, mlp_ratio=mlp_ratio, qkv_bias=qkv_bias,
qk_scale=qk_scale, drop=drop_rate, attn_drop=attn_drop_rate, drop_path=self.trans_dpr[0],
)
# 2~4 stage
init_stage = 2
fin_stage = depth // 3 + 1
for i in range(init_stage, fin_stage):
self.add_module('conv_trans_' + str(i),
ConvTransBlock(
stage_1_channel, stage_1_channel, False, 1, dw_stride=trans_dw_stride, embed_dim=embed_dim,
num_heads=num_heads, mlp_ratio=mlp_ratio, qkv_bias=qkv_bias, qk_scale=qk_scale,
drop_rate=drop_rate, attn_drop_rate=attn_drop_rate, drop_path_rate=self.trans_dpr[i-1],
num_med_block=num_med_block
)
)
stage_2_channel = int(base_channel * channel_ratio * 2)
# 5~8 stage
init_stage = fin_stage # 5
fin_stage = fin_stage + depth // 3 # 9
for i in range(init_stage, fin_stage):
s = 2 if i == init_stage else 1
in_channel = stage_1_channel if i == init_stage else stage_2_channel
res_conv = True if i == init_stage else False
self.add_module('conv_trans_' + str(i),
ConvTransBlock(
in_channel, stage_2_channel, res_conv, s, dw_stride=trans_dw_stride // 2, embed_dim=embed_dim,
num_heads=num_heads, mlp_ratio=mlp_ratio, qkv_bias=qkv_bias, qk_scale=qk_scale,
drop_rate=drop_rate, attn_drop_rate=attn_drop_rate, drop_path_rate=self.trans_dpr[i-1],
num_med_block=num_med_block
)
)
stage_3_channel = int(base_channel * channel_ratio * 2 * 2)
# 9~12 stage
init_stage = fin_stage # 9
fin_stage = fin_stage + depth // 3 # 13
for i in range(init_stage, fin_stage):
s = 2 if i == init_stage else 1
in_channel = stage_2_channel if i == init_stage else stage_3_channel
res_conv = True if i == init_stage else False
last_fusion = True if i == depth else False
self.add_module('conv_trans_' + str(i),
ConvTransBlock(
in_channel, stage_3_channel, res_conv, s, dw_stride=trans_dw_stride // 4, embed_dim=embed_dim,
num_heads=num_heads, mlp_ratio=mlp_ratio, qkv_bias=qkv_bias, qk_scale=qk_scale,
drop_rate=drop_rate, attn_drop_rate=attn_drop_rate, drop_path_rate=self.trans_dpr[i-1],
num_med_block=num_med_block, last_fusion=last_fusion
)
)
self.fin_stage = fin_stage
trunc_normal_(self.cls_token, std=.02)
self.apply(self._init_weights)
def _init_weights(self, m):
if isinstance(m, nn.Linear):
trunc_normal_(m.weight, std=.02)
if isinstance(m, nn.Linear) and m.bias is not None:
nn.init.constant_(m.bias, 0)
elif isinstance(m, nn.LayerNorm):
nn.init.constant_(m.bias, 0)
nn.init.constant_(m.weight, 1.0)
elif isinstance(m, nn.Conv2d):
nn.init.kaiming_normal_(m.weight, mode='fan_out', nonlinearity='relu')
elif isinstance(m, nn.BatchNorm2d):
nn.init.constant_(m.weight, 1.)
nn.init.constant_(m.bias, 0.)
elif isinstance(m, nn.GroupNorm):
nn.init.constant_(m.weight, 1.)
nn.init.constant_(m.bias, 0.)
@torch.jit.ignore
def no_weight_decay(self):
return {'cls_token'}
def forward(self, x):
B = x.shape[0]
cls_tokens = self.cls_token.expand(B, -1, -1)
# pdb.set_trace()
# stem stage [N, 3, 224, 224] -> [N, 64, 56, 56]
x_base = self.maxpool(self.act1(self.bn1(self.conv1(x))))
# 1 stage
x = self.conv_1(x_base, return_x_2=False)
x_t = self.trans_patch_conv(x_base).flatten(2).transpose(1, 2)
x_t = torch.cat([cls_tokens, x_t], dim=1)
x_t = self.trans_1(x_t)
# 2 ~ final
for i in range(2, self.fin_stage):
x, x_t = eval('self.conv_trans_' + str(i))(x, x_t)
# conv classification
x_p = self.pooling(x).flatten(1)
conv_cls = self.conv_cls_head(x_p)
# trans classification
x_t = self.trans_norm(x_t)
tran_cls = self.trans_cls_head(x_t[:, 0])
return [conv_cls, tran_cls]
PyTorch/NLP/Conformer-main/datasets.py 0 → 100644
View file @ a4372e17
import os
import json
from torchvision import datasets, transforms
from torchvision.datasets.folder import ImageFolder, default_loader
from timm.data.constants import IMAGENET_DEFAULT_MEAN, IMAGENET_DEFAULT_STD
from timm.data import create_transform
class INatDataset(ImageFolder):
def __init__(self, root, train=True, year=2018, transform=None, target_transform=None,
category='name', loader=default_loader):
self.transform = transform
self.loader = loader
self.target_transform = target_transform
self.year = year
# assert category in ['kingdom','phylum','class','order','supercategory','family','genus','name']
path_json = os.path.join(root, f'{"train" if train else "val"}{year}.json')
with open(path_json) as json_file:
data = json.load(json_file)
with open(os.path.join(root, 'categories.json')) as json_file:
data_catg = json.load(json_file)
path_json_for_targeter = os.path.join(root, f"train{year}.json")
with open(path_json_for_targeter) as json_file:
data_for_targeter = json.load(json_file)
targeter = {}
indexer = 0
for elem in data_for_targeter['annotations']:
king = []
king.append(data_catg[int(elem['category_id'])][category])
if king[0] not in targeter.keys():
targeter[king[0]] = indexer
indexer += 1
self.nb_classes = len(targeter)
self.samples = []
for elem in data['images']:
cut = elem['file_name'].split('/')
target_current = int(cut[2])
path_current = os.path.join(root, cut[0], cut[2], cut[3])
categors = data_catg[target_current]
target_current_true = targeter[categors[category]]
self.samples.append((path_current, target_current_true))
# __getitem__ and __len__ inherited from ImageFolder
def build_dataset(is_train, args):
transform = build_transform(is_train, args)
if args.data_set == 'CIFAR':
dataset = datasets.CIFAR100(args.data_path, train=is_train, transform=transform)
nb_classes = 100
elif args.data_set == 'CIFAR10':
dataset = datasets.CIFAR10(args.data_path, train=is_train, transform=transform)
nb_classes = 10
elif args.data_set == 'IMNET':
root = os.path.join(args.data_path, 'train' if is_train else 'val')
dataset = datasets.ImageFolder(root, transform=transform)
nb_classes = 1000
elif args.data_set == 'INAT':
dataset = INatDataset(args.data_path, train=is_train, year=2018,
category=args.inat_category, transform=transform)
nb_classes = dataset.nb_classes
elif args.data_set == 'INAT19':
dataset = INatDataset(args.data_path, train=is_train, year=2019,
category=args.inat_category, transform=transform)
nb_classes = dataset.nb_classes
return dataset, nb_classes
def build_transform(is_train, args):
resize_im = args.input_size > 32
if is_train:
# this should always dispatch to transforms_imagenet_train
transform = create_transform(
input_size=args.input_size,
is_training=True,
color_jitter=args.color_jitter,
auto_augment=args.aa,
interpolation=args.train_interpolation,
re_prob=args.reprob,
re_mode=args.remode,
re_count=args.recount,
)
if not resize_im:
# replace RandomResizedCropAndInterpolation with
# RandomCrop
transform.transforms[0] = transforms.RandomCrop(
args.input_size, padding=4)
return transform
t = []
if resize_im:
size = int((256 / 224) * args.input_size)
t.append(
transforms.Resize(size, interpolation=3), # to maintain same ratio w.r.t. 224 images
)
t.append(transforms.CenterCrop(args.input_size))
t.append(transforms.ToTensor())
t.append(transforms.Normalize(IMAGENET_DEFAULT_MEAN, IMAGENET_DEFAULT_STD))
return transforms.Compose(t)
PyTorch/NLP/Conformer-main/engine.py 0 → 100644
View file @ a4372e17
"""
Train and eval functions used in main.py
"""
import math
import sys
from typing import Iterable, Optional
import torch
from timm.data import Mixup
from timm.utils import accuracy, ModelEma
import utils
def train_one_epoch(model: torch.nn.Module, criterion: torch.nn.Module,
data_loader: Iterable, optimizer: torch.optim.Optimizer,
device: torch.device, epoch: int, loss_scaler, max_norm: float = 0,
model_ema: Optional[ModelEma] = None, mixup_fn: Optional[Mixup] = None,
set_training_mode=True
):
# TODO fix this for finetuning
model.train(set_training_mode)
criterion.train()
metric_logger = utils.MetricLogger(delimiter=" ")
metric_logger.add_meter('lr', utils.SmoothedValue(window_size=1, fmt='{value:.6f}'))
header = 'Epoch: [{}]'.format(epoch)
print_freq = 10
for samples, targets in metric_logger.log_every(data_loader, print_freq, header):
samples = samples.to(device, non_blocking=True)
targets = targets.to(device, non_blocking=True)
if mixup_fn is not None:
samples, targets = mixup_fn(samples, targets)
with torch.cuda.amp.autocast():
outputs = model(samples)
if isinstance(outputs, list):
loss_list = [criterion(o, targets) / len(outputs) for o in outputs]
loss = sum(loss_list)
else:
loss = criterion(outputs, targets)
loss_value = loss.item()
if not math.isfinite(loss_value):
print("Loss is {}, stopping training".format(loss_value))
sys.exit(1)
optimizer.zero_grad()
# this attribute is added by timm on one optimizer (adahessian)
is_second_order = hasattr(optimizer, 'is_second_order') and optimizer.is_second_order
loss_scaler(loss, optimizer, clip_grad=max_norm,
parameters=model.parameters(), create_graph=is_second_order)
torch.cuda.synchronize()
if model_ema is not None:
model_ema.update(model)
if isinstance(outputs, list):
metric_logger.update(loss_0=loss_list[0].item())
metric_logger.update(loss_1=loss_list[1].item())
else:
metric_logger.update(loss=loss_value)
metric_logger.update(lr=optimizer.param_groups[0]["lr"])
# gather the stats from all processes
metric_logger.synchronize_between_processes()
print("Averaged stats:", metric_logger)
return {k: meter.global_avg for k, meter in metric_logger.meters.items()}
@torch.no_grad()
def evaluate(data_loader, model, device):
criterion = torch.nn.CrossEntropyLoss()
metric_logger = utils.MetricLogger(delimiter=" ")
header = 'Test:'
# switch to evaluation mode
model.eval()
for images, target in metric_logger.log_every(data_loader, 10, header):
images = images.to(device, non_blocking=True)
target = target.to(device, non_blocking=True)
# compute output
with torch.cuda.amp.autocast():
output = model(images)
# Conformer
if isinstance(output, list):
loss_list = [criterion(o, target) / len(output) for o in output]
loss = sum(loss_list)
# others
else:
loss = criterion(output, target)
if isinstance(output, list):
# Conformer
acc1_head1 = accuracy(output[0], target, topk=(1,))[0]
acc1_head2 = accuracy(output[1], target, topk=(1,))[0]
acc1_total = accuracy(output[0] + output[1], target, topk=(1,))[0]
else:
# others
acc1, acc5 = accuracy(output, target, topk=(1, 5))
batch_size = images.shape[0]
if isinstance(output, list):
metric_logger.update(loss=loss.item())
metric_logger.update(loss_0=loss_list[0].item())
metric_logger.update(loss_1=loss_list[1].item())
metric_logger.meters['acc1'].update(acc1_total.item(), n=batch_size)
metric_logger.meters['acc1_head1'].update(acc1_head1.item(), n=batch_size)
metric_logger.meters['acc1_head2'].update(acc1_head2.item(), n=batch_size)
else:
metric_logger.update(loss=loss.item())
metric_logger.meters['acc1'].update(acc1.item(), n=batch_size)
metric_logger.meters['acc5'].update(acc5.item(), n=batch_size)
if isinstance(output, list):
print('* Acc@heads_top1 {heads_top1.global_avg:.3f} Acc@head_1 {head1_top1.global_avg:.3f} Acc@head_2 {head2_top1.global_avg:.3f} '
'loss@total {losses.global_avg:.3f} loss@1 {loss_0.global_avg:.3f} loss@2 {loss_1.global_avg:.3f} '
.format(heads_top1=metric_logger.acc1, head1_top1=metric_logger.acc1_head1, head2_top1=metric_logger.acc1_head2,
losses=metric_logger.loss, loss_0=metric_logger.loss_0, loss_1=metric_logger.loss_1))
else:
print('* Acc@1 {top1.global_avg:.3f} Acc@5 {top5.global_avg:.3f} loss {losses.global_avg:.3f}'
.format(top1=metric_logger.acc1, top5=metric_logger.acc5, losses=metric_logger.loss))
return {k: meter.global_avg for k, meter in metric_logger.meters.items()}
PyTorch/NLP/Conformer-main/hubconf.py 0 → 100644
View file @ a4372e17
dependencies = ["torch", "torchvision", "timm"]
PyTorch/NLP/Conformer-main/main.py 0 → 100644
View file @ a4372e17
import argparse
import datetime
import numpy as np
import time
import torch
import torch.backends.cudnn as cudnn
import json
from pathlib import Path
from timm.data import Mixup
from timm.models import create_model
from timm.loss import LabelSmoothingCrossEntropy, SoftTargetCrossEntropy
from timm.scheduler import create_scheduler
from timm.optim import create_optimizer
from timm.utils import NativeScaler, get_state_dict, ModelEma
from datasets import build_dataset
from engine import train_one_epoch, evaluate
from samplers import RASampler
import utils
import models
# from fvcore.nn import FlopCountAnalysis
def get_args_parser():
parser = argparse.ArgumentParser('DeiT training and evaluation script', add_help=False)
parser.add_argument('--batch-size', default=64, type=int)
parser.add_argument('--epochs', default=300, type=int)
# Model parameters
parser.add_argument('--model', default='deit_base_patch16_224', type=str, metavar='MODEL',
help='Name of model to train')
parser.add_argument('--input-size', default=224, type=int, help='images input size')
parser.add_argument('--drop', type=float, default=0.0, metavar='PCT',
help='Dropout rate (default: 0.)')
parser.add_argument('--drop-path', type=float, default=0.1, metavar='PCT',
help='Drop path rate (default: 0.1)')
parser.add_argument('--drop-block', type=float, default=None, metavar='PCT',
help='Drop block rate (default: None)')
parser.add_argument('--model-ema', action='store_true')
parser.add_argument('--no-model-ema', action='store_false', dest='model_ema')
parser.set_defaults(model_ema=True)
parser.add_argument('--model-ema-decay', type=float, default=0.99996, help='')
parser.add_argument('--model-ema-force-cpu', action='store_true', default=False, help='')
# Optimizer parameters
parser.add_argument('--opt', default='adamw', type=str, metavar='OPTIMIZER',
help='Optimizer (default: "adamw"')
parser.add_argument('--opt-eps', default=1e-8, type=float, metavar='EPSILON',
help='Optimizer Epsilon (default: 1e-8)')
parser.add_argument('--opt-betas', default=None, type=float, nargs='+', metavar='BETA',
help='Optimizer Betas (default: None, use opt default)')
parser.add_argument('--clip-grad', type=float, default=None, metavar='NORM',
help='Clip gradient norm (default: None, no clipping)')
parser.add_argument('--momentum', type=float, default=0.9, metavar='M',
help='SGD momentum (default: 0.9)')
parser.add_argument('--weight-decay', type=float, default=0.05,
help='weight decay (default: 0.05)')
# Learning rate schedule parameters
parser.add_argument('--sched', default='cosine', type=str, metavar='SCHEDULER',
help='LR scheduler (default: "cosine"')
parser.add_argument('--lr', type=float, default=5e-4, metavar='LR',
help='learning rate (default: 5e-4)')
parser.add_argument('--lr-noise', type=float, nargs='+', default=None, metavar='pct, pct',
help='learning rate noise on/off epoch percentages')
parser.add_argument('--lr-noise-pct', type=float, default=0.67, metavar='PERCENT',
help='learning rate noise limit percent (default: 0.67)')
parser.add_argument('--lr-noise-std', type=float, default=1.0, metavar='STDDEV',
help='learning rate noise std-dev (default: 1.0)')
parser.add_argument('--warmup-lr', type=float, default=1e-6, metavar='LR',
help='warmup learning rate (default: 1e-6)')
parser.add_argument('--min-lr', type=float, default=1e-5, metavar='LR',
help='lower lr bound for cyclic schedulers that hit 0 (1e-5)')
parser.add_argument('--decay-epochs', type=float, default=30, metavar='N',
help='epoch interval to decay LR')
parser.add_argument('--warmup-epochs', type=int, default=5, metavar='N',
help='epochs to warmup LR, if scheduler supports')
parser.add_argument('--cooldown-epochs', type=int, default=10, metavar='N',
help='epochs to cooldown LR at min_lr, after cyclic schedule ends')
parser.add_argument('--patience-epochs', type=int, default=10, metavar='N',
help='patience epochs for Plateau LR scheduler (default: 10')
parser.add_argument('--decay-rate', '--dr', type=float, default=0.1, metavar='RATE',
help='LR decay rate (default: 0.1)')
# Augmentation parameters
parser.add_argument('--color-jitter', type=float, default=0.4, metavar='PCT',
help='Color jitter factor (default: 0.4)')
parser.add_argument('--aa', type=str, default='rand-m9-mstd0.5-inc1', metavar='NAME',
help='Use AutoAugment policy. "v0" or "original". " + \
"(default: rand-m9-mstd0.5-inc1)'),
parser.add_argument('--smoothing', type=float, default=0.1, help='Label smoothing (default: 0.1)')
parser.add_argument('--train-interpolation', type=str, default='bicubic',
help='Training interpolation (random, bilinear, bicubic default: "bicubic")')
parser.add_argument('--repeated-aug', action='store_true')
parser.add_argument('--no-repeated-aug', action='store_false', dest='repeated_aug')
parser.set_defaults(repeated_aug=True)
# * Random Erase params
parser.add_argument('--reprob', type=float, default=0.25, metavar='PCT',
help='Random erase prob (default: 0.25)')
parser.add_argument('--remode', type=str, default='pixel',
help='Random erase mode (default: "pixel")')
parser.add_argument('--recount', type=int, default=1,
help='Random erase count (default: 1)')
parser.add_argument('--resplit', action='store_true', default=False,
help='Do not random erase first (clean) augmentation split')
# * Mixup params
parser.add_argument('--mixup', type=float, default=0.8,
help='mixup alpha, mixup enabled if > 0. (default: 0.8)')
parser.add_argument('--cutmix', type=float, default=1.0,
help='cutmix alpha, cutmix enabled if > 0. (default: 1.0)')
parser.add_argument('--cutmix-minmax', type=float, nargs='+', default=None,
help='cutmix min/max ratio, overrides alpha and enables cutmix if set (default: None)')
parser.add_argument('--mixup-prob', type=float, default=1.0,
help='Probability of performing mixup or cutmix when either/both is enabled')
parser.add_argument('--mixup-switch-prob', type=float, default=0.5,
help='Probability of switching to cutmix when both mixup and cutmix enabled')
parser.add_argument('--mixup-mode', type=str, default='batch',
help='How to apply mixup/cutmix params. Per "batch", "pair", or "elem"')
# Dataset parameters
parser.add_argument('--data-path', default='/datasets01_101/imagenet_full_size/061417/', type=str,
help='dataset path')
parser.add_argument('--data-set', default='IMNET', choices=['CIFAR', 'CIFAR10', 'IMNET', 'INAT', 'INAT19'],
type=str, help='Image Net dataset path')
parser.add_argument('--inat-category', default='name',
choices=['kingdom', 'phylum', 'class', 'order', 'supercategory', 'family', 'genus', 'name'],
type=str, help='semantic granularity')
# * Finetuning params
parser.add_argument('--finetune', default='', help='finetune from checkpoint')
parser.add_argument('--evaluate-freq', type=int, default=1, help='frequency of perform evaluation (default: 5)')
parser.add_argument('--output_dir', default='',
help='path where to save, empty for no saving')
parser.add_argument('--device', default='cuda',
help='device to use for training / testing')
parser.add_argument('--seed', default=0, type=int)
parser.add_argument('--resume', default='', help='resume from checkpoint')
parser.add_argument('--start_epoch', default=0, type=int, metavar='N',
help='start epoch')
parser.add_argument('--eval', action='store_true', help='Perform evaluation only')
parser.add_argument('--num_workers', default=10, type=int)
parser.add_argument('--pin-mem', action='store_true',
help='Pin CPU memory in DataLoader for more efficient (sometimes) transfer to GPU.')
parser.add_argument('--no-pin-mem', action='store_false', dest='pin_mem',
help='')
parser.set_defaults(pin_mem=True)
# distributed training parameters
parser.add_argument('--world_size', default=1, type=int,
help='number of distributed processes')
parser.add_argument('--dist_url', default='env://', help='url used to set up distributed training')
return parser
def main(args):
utils.init_distributed_mode(args)
print(args)
device = torch.device(args.device)
# fix the seed for reproducibility
seed = args.seed + utils.get_rank()
torch.manual_seed(seed)
np.random.seed(seed)
# random.seed(seed)
cudnn.benchmark = True
dataset_train, args.nb_classes = build_dataset(is_train=True, args=args)
dataset_val, _ = build_dataset(is_train=False, args=args)
if True: # args.distributed:
num_tasks = utils.get_world_size()
global_rank = utils.get_rank()
if args.repeated_aug:
sampler_train = RASampler(
dataset_train, num_replicas=num_tasks, rank=global_rank, shuffle=True
)
else:
sampler_train = torch.utils.data.DistributedSampler(
dataset_train, num_replicas=num_tasks, rank=global_rank, shuffle=True
)
else:
sampler_train = torch.utils.data.RandomSampler(dataset_train)
data_loader_train = torch.utils.data.DataLoader(
dataset_train, sampler=sampler_train,
batch_size=args.batch_size,
num_workers=args.num_workers,
pin_memory=args.pin_mem,
drop_last=True,
)
data_loader_val = torch.utils.data.DataLoader(
dataset_val, batch_size=int(3.0 * args.batch_size),
shuffle=False, num_workers=args.num_workers,
pin_memory=args.pin_mem, drop_last=False
)
mixup_fn = None
mixup_active = args.mixup > 0 or args.cutmix > 0. or args.cutmix_minmax is not None
if mixup_active:
mixup_fn = Mixup(
mixup_alpha=args.mixup, cutmix_alpha=args.cutmix, cutmix_minmax=args.cutmix_minmax,
prob=args.mixup_prob, switch_prob=args.mixup_switch_prob, mode=args.mixup_mode,
label_smoothing=args.smoothing, num_classes=args.nb_classes)
print(f"Creating model: {args.model}")
model = create_model(
args.model,
pretrained=False,
num_classes=args.nb_classes,
drop_rate=args.drop,
drop_path_rate=args.drop_path,
drop_block_rate=args.drop_block,
)
if args.finetune:
if args.finetune.startswith('https'):
checkpoint = torch.hub.load_state_dict_from_url(
args.finetune, map_location='cpu', check_hash=True)
else:
checkpoint = torch.load(args.finetune, map_location='cpu')
if 'model' in checkpoint.keys():
checkpoint_model = checkpoint['model']
else:
checkpoint_model = checkpoint
state_dict = model.state_dict()
for k in ['head.weight', 'head.bias', 'head_dist.weight', 'head_dist.bias',
'trans_cls_head.weight', 'trans_cls_head.bias', 'conv_cls_head.weight', 'conv_cls_head.bias']:
if k in checkpoint_model and checkpoint_model[k].shape != state_dict[k].shape:
print(f"Removing key {k} from pretrained checkpoint")
del checkpoint_model[k]
if 'pos_embed' in checkpoint_model.keys():
# interpolate position embedding
pos_embed_checkpoint = checkpoint_model['pos_embed']
embedding_size = pos_embed_checkpoint.shape[-1]
num_patches = model.patch_embed.num_patches
num_extra_tokens = model.pos_embed.shape[-2] - num_patches
# height (== width) for the checkpoint position embedding
orig_size = int((pos_embed_checkpoint.shape[-2] - num_extra_tokens) ** 0.5)
# height (== width) for the new position embedding
new_size = int(num_patches ** 0.5)
# class_token and dist_token are kept unchanged
extra_tokens = pos_embed_checkpoint[:, :num_extra_tokens]
# only the position tokens are interpolated
pos_tokens = pos_embed_checkpoint[:, num_extra_tokens:]
pos_tokens = pos_tokens.reshape(-1, orig_size, orig_size, embedding_size).permute(0, 3, 1, 2)
pos_tokens = torch.nn.functional.interpolate(
pos_tokens, size=(new_size, new_size), mode='bicubic', align_corners=False)
pos_tokens = pos_tokens.permute(0, 2, 3, 1).flatten(1, 2)
new_pos_embed = torch.cat((extra_tokens, pos_tokens), dim=1)
checkpoint_model['pos_embed'] = new_pos_embed
model.load_state_dict(checkpoint_model, strict=False)
model.to(device)
model_ema = None
if args.model_ema:
# Important to create EMA model after cuda(), DP wrapper, and AMP but before SyncBN and DDP wrapper
model_ema = ModelEma(
model,
decay=args.model_ema_decay,
device='cpu' if args.model_ema_force_cpu else '',
resume='')
model_without_ddp = model
if args.distributed:
model = torch.nn.parallel.DistributedDataParallel(model, device_ids=[args.gpu])
model_without_ddp = model.module
n_parameters = sum(p.numel() for p in model.parameters() if p.requires_grad)
print('number of params:', n_parameters)
# linear_scaled_lr = args.lr * args.batch_size * utils.get_world_size() / 512.0
# args.lr = linear_scaled_lr
optimizer = create_optimizer(args, model)
loss_scaler = NativeScaler()
lr_scheduler, _ = create_scheduler(args, optimizer)
criterion = LabelSmoothingCrossEntropy()
if args.mixup > 0.:
# smoothing is handled with mixup label transform
criterion = SoftTargetCrossEntropy()
elif args.smoothing:
criterion = LabelSmoothingCrossEntropy(smoothing=args.smoothing)
else:
criterion = torch.nn.CrossEntropyLoss()
output_dir = Path(args.output_dir)
if args.resume:
if args.resume.startswith('https'):
checkpoint = torch.hub.load_state_dict_from_url(
args.resume, map_location='cpu', check_hash=True)
else:
checkpoint = torch.load(args.resume, map_location='cpu')
# pdb.set_trace()
if 'model' in checkpoint.keys():
model_without_ddp.load_state_dict(checkpoint['model'])
else:
model_without_ddp.load_state_dict(checkpoint)
if not args.eval and 'optimizer' in checkpoint and 'lr_scheduler' in checkpoint and 'epoch' in checkpoint:
optimizer.load_state_dict(checkpoint['optimizer'])
lr_scheduler.load_state_dict(checkpoint['lr_scheduler'])
args.start_epoch = checkpoint['epoch'] + 1
if args.model_ema:
utils._load_checkpoint_for_ema(model_ema, checkpoint['model_ema'])
if args.eval:
test_stats = evaluate(data_loader_val, model, device)
print(f"Accuracy of the network on the {len(dataset_val)} test images: {test_stats['acc1']:.1f}%")
return
print("Start training")
start_time = time.time()
max_accuracy = 0.0
for epoch in range(args.start_epoch, args.epochs):
if args.distributed:
data_loader_train.sampler.set_epoch(epoch)
train_stats = train_one_epoch(
model, criterion, data_loader_train,
optimizer, device, epoch, loss_scaler,
args.clip_grad, model_ema, mixup_fn,
set_training_mode=args.finetune == '' # keep in eval mode during finetuning
)
lr_scheduler.step(epoch)
if args.output_dir:
checkpoint_paths = [output_dir / 'checkpoint.pth']
for checkpoint_path in checkpoint_paths:
utils.save_on_master({
'model': model_without_ddp.state_dict(),
'optimizer': optimizer.state_dict(),
'lr_scheduler': lr_scheduler.state_dict(),
'epoch': epoch,
'model_ema': get_state_dict(model_ema),
'args': args,
}, checkpoint_path)
if epoch % args.evaluate_freq == 0:
test_stats = evaluate(data_loader_val, model, device)
print(f"Accuracy of the network on the {len(dataset_val)} test images: {test_stats['acc1']:.1f}%")
max_accuracy = max(max_accuracy, test_stats["acc1"])
print(f'Max accuracy: {max_accuracy:.2f}%')
log_stats = {**{f'train_{k}': v for k, v in train_stats.items()},
**{f'test_{k}': v for k, v in test_stats.items()},
'epoch': epoch,
'n_parameters': n_parameters}
if args.output_dir and utils.is_main_process():
with (output_dir / "log.txt").open("a") as f:
f.write(json.dumps(log_stats) + "\n")
total_time = time.time() - start_time
total_time_str = str(datetime.timedelta(seconds=int(total_time)))
print('Training time {}'.format(total_time_str))
if __name__ == '__main__':
parser = argparse.ArgumentParser('DeiT training and evaluation script', parents=[get_args_parser()])
args = parser.parse_args()
if args.output_dir:
Path(args.output_dir).mkdir(parents=True, exist_ok=True)
main(args)
PyTorch/NLP/Conformer-main/mmdetection/.dev_scripts/batch_test.py 0 → 100644
View file @ a4372e17
"""
some instructions
1. Fill the models that needs to be checked in the modelzoo_dict
2. Arange the structure of the directory as follows, the script will find the
corresponding config itself:
model_dir/model_family/checkpoints
e.g.: models/faster_rcnn/faster_rcnn_r50_fpn_1x_coco_20200130-047c8118.pth
models/faster_rcnn/faster_rcnn_r101_fpn_1x_coco_20200130-047c8118.pth
3. Excute the batch_test.sh
"""
import argparse
import json
import os
import subprocess
import mmcv
import torch
from mmcv import Config, get_logger
from mmcv.parallel import MMDataParallel, MMDistributedDataParallel
from mmcv.runner import (get_dist_info, init_dist, load_checkpoint,
wrap_fp16_model)
from mmdet.apis import multi_gpu_test, single_gpu_test
from mmdet.datasets import (build_dataloader, build_dataset,
replace_ImageToTensor)
from mmdet.models import build_detector
modelzoo_dict = {
'configs/faster_rcnn/faster_rcnn_r50_fpn_1x_coco.py': {
'bbox': 0.374
},
'configs/mask_rcnn/mask_rcnn_r50_fpn_1x_coco.py': {
'bbox': 0.382,
'segm': 0.347
},
'configs/rpn/rpn_r50_fpn_1x_coco.py': {
'AR@1000': 0.582
}
}
def parse_args():
parser = argparse.ArgumentParser(
description='The script used for checking the correctness \
of batch inference')
parser.add_argument('model_dir', help='directory of models')
parser.add_argument(
'json_out', help='the output json records test information like mAP')
parser.add_argument(
'--launcher',
choices=['none', 'pytorch', 'slurm', 'mpi'],
default='none',
help='job launcher')
parser.add_argument('--local_rank', type=int, default=0)
args = parser.parse_args()
if 'LOCAL_RANK' not in os.environ:
os.environ['LOCAL_RANK'] = str(args.local_rank)
return args
def check_finish(all_model_dict, result_file):
# check if all models are checked
tested_cfgs = []
with open(result_file, 'r+') as f:
for line in f:
line = json.loads(line)
tested_cfgs.append(line['cfg'])
is_finish = True
for cfg in sorted(all_model_dict.keys()):
if cfg not in tested_cfgs:
return cfg
if is_finish:
with open(result_file, 'a+') as f:
f.write('finished\n')
def dump_dict(record_dict, json_out):
# dump result json dict
with open(json_out, 'a+') as f:
mmcv.dump(record_dict, f, file_format='json')
f.write('\n')
def main():
args = parse_args()
# touch the output json if not exist
with open(args.json_out, 'a+'):
pass
# init distributed env first, since logger depends on the dist
# info.
if args.launcher == 'none':
distributed = False
else:
distributed = True
init_dist(args.launcher, backend='nccl')
rank, world_size = get_dist_info()
logger = get_logger('root')
# read info of checkpoints and config
result_dict = dict()
for model_family_dir in os.listdir(args.model_dir):
for model in os.listdir(
os.path.join(args.model_dir, model_family_dir)):
# cpt: rpn_r50_fpn_1x_coco_20200218-5525fa2e.pth
# cfg: rpn_r50_fpn_1x_coco.py
cfg = model.split('.')[0][:-18] + '.py'
cfg_path = os.path.join('configs', model_family_dir, cfg)
assert os.path.isfile(
cfg_path), f'{cfg_path} is not valid config path'
cpt_path = os.path.join(args.model_dir, model_family_dir, model)
result_dict[cfg_path] = cpt_path
assert cfg_path in modelzoo_dict, f'please fill the ' \
f'performance of cfg: {cfg_path}'
cfg = check_finish(result_dict, args.json_out)
cpt = result_dict[cfg]
try:
cfg_name = cfg
logger.info(f'evaluate {cfg}')
record = dict(cfg=cfg, cpt=cpt)
cfg = Config.fromfile(cfg)
# cfg.data.test.ann_file = 'data/val_0_10.json'
# set cudnn_benchmark
if cfg.get('cudnn_benchmark', False):
torch.backends.cudnn.benchmark = True
cfg.model.pretrained = None
if cfg.model.get('neck'):
if isinstance(cfg.model.neck, list):
for neck_cfg in cfg.model.neck:
if neck_cfg.get('rfp_backbone'):
if neck_cfg.rfp_backbone.get('pretrained'):
neck_cfg.rfp_backbone.pretrained = None
elif cfg.model.neck.get('rfp_backbone'):
if cfg.model.neck.rfp_backbone.get('pretrained'):
cfg.model.neck.rfp_backbone.pretrained = None
# in case the test dataset is concatenated
if isinstance(cfg.data.test, dict):
cfg.data.test.test_mode = True
elif isinstance(cfg.data.test, list):
for ds_cfg in cfg.data.test:
ds_cfg.test_mode = True
# build the dataloader
samples_per_gpu = 2 # hack test with 2 image per gpu
if samples_per_gpu > 1:
# Replace 'ImageToTensor' to 'DefaultFormatBundle'
cfg.data.test.pipeline = replace_ImageToTensor(
cfg.data.test.pipeline)
dataset = build_dataset(cfg.data.test)
data_loader = build_dataloader(
dataset,
samples_per_gpu=samples_per_gpu,
workers_per_gpu=cfg.data.workers_per_gpu,
dist=distributed,
shuffle=False)
# build the model and load checkpoint
cfg.model.train_cfg = None
model = build_detector(cfg.model, test_cfg=cfg.get('test_cfg'))
fp16_cfg = cfg.get('fp16', None)
if fp16_cfg is not None:
wrap_fp16_model(model)
checkpoint = load_checkpoint(model, cpt, map_location='cpu')
# old versions did not save class info in checkpoints,
# this walkaround is for backward compatibility
if 'CLASSES' in checkpoint['meta']:
model.CLASSES = checkpoint['meta']['CLASSES']
else:
model.CLASSES = dataset.CLASSES
if not distributed:
model = MMDataParallel(model, device_ids=[0])
outputs = single_gpu_test(model, data_loader)
else:
model = MMDistributedDataParallel(
model.cuda(),
device_ids=[torch.cuda.current_device()],
broadcast_buffers=False)
outputs = multi_gpu_test(model, data_loader, 'tmp')
if rank == 0:
ref_mAP_dict = modelzoo_dict[cfg_name]
metrics = list(ref_mAP_dict.keys())
metrics = [
m if m != 'AR@1000' else 'proposal_fast' for m in metrics
]
eval_results = dataset.evaluate(outputs, metrics)
print(eval_results)
for metric in metrics:
if metric == 'proposal_fast':
ref_metric = modelzoo_dict[cfg_name]['AR@1000']
eval_metric = eval_results['AR@1000']
else:
ref_metric = modelzoo_dict[cfg_name][metric]
eval_metric = eval_results[f'{metric}_mAP']
if abs(ref_metric - eval_metric) > 0.003:
record['is_normal'] = False
dump_dict(record, args.json_out)
check_finish(result_dict, args.json_out)
except Exception as e:
logger.error(f'rank: {rank} test fail with error: {e}')
record['terminate'] = True
dump_dict(record, args.json_out)
check_finish(result_dict, args.json_out)
# hack there to throw some error to prevent hang out
subprocess.call('xxx')
if __name__ == '__main__':
main()
PyTorch/NLP/Conformer-main/mmdetection/.dev_scripts/batch_test.sh 0 → 100644
View file @ a4372e17
export PYTHONPATH=${PWD}
partition=$1
model_dir=$2
json_out=$3
job_name=batch_test
gpus=8
gpu_per_node=8
touch $json_out
lastLine=$(tail -n 1 $json_out)
while [ "$lastLine" != "finished" ]
do
srun -p ${partition} --gres=gpu:${gpu_per_node} -n${gpus} --ntasks-per-node=${gpu_per_node} \
--job-name=${job_name} --kill-on-bad-exit=1 \
python .dev_scripts/batch_test.py $model_dir $json_out --launcher='slurm'
lastLine=$(tail -n 1 $json_out)
echo $lastLine
done
PyTorch/NLP/Conformer-main/mmdetection/.dev_scripts/benchmark_filter.py 0 → 100644
View file @ a4372e17
import argparse
import os
import os.path as osp
import mmcv
def parse_args():
parser = argparse.ArgumentParser(description='Filter configs to train')
parser.add_argument(
'--basic-arch',
action='store_true',
help='to train models in basic arch')
parser.add_argument(
'--datasets', action='store_true', help='to train models in dataset')
parser.add_argument(
'--data-pipeline',
action='store_true',
help='to train models related to data pipeline, e.g. augmentations')
parser.add_argument(
'--nn-module',
action='store_true',
help='to train models related to neural network modules')
parser.add_argument(
'--model-options',
nargs='+',
help='custom options to special model benchmark')
args = parser.parse_args()
return args
basic_arch_root = [
'atss', 'cascade_rcnn', 'cascade_rpn', 'centripetalnet', 'cornernet',
'detectors', 'detr', 'double_heads', 'dynamic_rcnn', 'faster_rcnn', 'fcos',
'foveabox', 'fp16', 'free_anchor', 'fsaf', 'gfl', 'ghm', 'grid_rcnn',
'guided_anchoring', 'htc', 'libra_rcnn', 'mask_rcnn', 'ms_rcnn',
'nas_fcos', 'paa', 'pisa', 'point_rend', 'reppoints', 'retinanet', 'rpn',
'sabl', 'ssd', 'tridentnet', 'vfnet', 'yolact', 'yolo'
]
datasets_root = [
'wider_face', 'pascal_voc', 'cityscapes', 'lvis', 'deepfashion'
]
data_pipeline_root = ['albu_example', 'instaboost']
nn_module_root = [
'carafe', 'dcn', 'empirical_attention', 'gcnet', 'gn', 'gn+ws', 'hrnet',
'pafpn', 'nas_fpn', 'regnet', 'resnest', 'res2net', 'groie'
]
benchmark_pool = [
'configs/albu_example/mask_rcnn_r50_fpn_albu_1x_coco.py',
'configs/atss/atss_r50_fpn_1x_coco.py',
'configs/carafe/mask_rcnn_r50_fpn_carafe_1x_coco.py',
'configs/cascade_rcnn/cascade_mask_rcnn_r50_fpn_1x_coco.py',
'configs/cascade_rpn/crpn_faster_rcnn_r50_caffe_fpn_1x_coco.py',
'configs/centripetalnet/'
'centripetalnet_hourglass104_mstest_16x6_210e_coco.py',
'configs/cityscapes/mask_rcnn_r50_fpn_1x_cityscapes.py',
'configs/cornernet/'
'cornernet_hourglass104_mstest_8x6_210e_coco.py', # special
'configs/dcn/mask_rcnn_r50_fpn_mdconv_c3-c5_1x_coco.py',
'configs/dcn/faster_rcnn_r50_fpn_dpool_1x_coco.py',
'configs/dcn/faster_rcnn_r50_fpn_mdpool_1x_coco.py',
'configs/dcn/mask_rcnn_r50_fpn_dconv_c3-c5_1x_coco.py',
'configs/detectors/detectors_htc_r50_1x_coco.py',
'configs/detr/detr_r50_8x2_150e_coco.py',
'configs/double_heads/dh_faster_rcnn_r50_fpn_1x_coco.py',
'configs/dynamic_rcnn/dynamic_rcnn_r50_fpn_1x.py',
'configs/empirical_attention/faster_rcnn_r50_fpn_attention_1111_dcn_1x_coco.py', # noqa
'configs/faster_rcnn/faster_rcnn_r50_fpn_1x_coco.py',
'configs/faster_rcnn/faster_rcnn_r50_fpn_ohem_1x_coco.py',
'configs/faster_rcnn/faster_rcnn_r50_caffe_fpn_1x_coco.py',
'configs/faster_rcnn/faster_rcnn_r50_caffe_fpn_mstrain_1x_coco.py',
'configs/faster_rcnn/faster_rcnn_r50_caffe_dc5_mstrain_1x_coco.py',
'configs/fcos/fcos_center_r50_caffe_fpn_gn-head_4x4_1x_coco.py',
'configs/foveabox/fovea_align_r50_fpn_gn-head_4x4_2x_coco.py',
'configs/fp16/retinanet_r50_fpn_fp16_1x_coco.py',
'configs/fp16/mask_rcnn_r50_fpn_fp16_1x_coco.py',
'configs/free_anchor/retinanet_free_anchor_r50_fpn_1x_coco.py',
'configs/fsaf/fsaf_r50_fpn_1x_coco.py',
'configs/gcnet/mask_rcnn_r50_fpn_r4_gcb_c3-c5_1x_coco.py',
'configs/gfl/gfl_r50_fpn_1x_coco.py',
'configs/ghm/retinanet_ghm_r50_fpn_1x_coco.py',
'configs/gn/mask_rcnn_r50_fpn_gn-all_2x_coco.py',
'configs/gn+ws/mask_rcnn_r50_fpn_gn_ws-all_2x_coco.py',
'configs/grid_rcnn/grid_rcnn_r50_fpn_gn-head_2x_coco.py',
'configs/groie/faster_rcnn_r50_fpn_groie_1x_coco.py',
'configs/guided_anchoring/ga_faster_r50_caffe_fpn_1x_coco.py',
'configs/hrnet/mask_rcnn_hrnetv2p_w18_1x_coco.py',
'configs/htc/htc_r50_fpn_1x_coco.py',
'configs/instaboost/mask_rcnn_r50_fpn_instaboost_4x_coco.py',
'configs/libra_rcnn/libra_faster_rcnn_r50_fpn_1x_coco.py',
'configs/lvis/mask_rcnn_r50_fpn_sample1e-3_mstrain_1x_lvis_v1.py',
'configs/mask_rcnn/mask_rcnn_r50_caffe_fpn_mstrain-poly_1x_coco.py',
'configs/ms_rcnn/ms_rcnn_r50_caffe_fpn_1x_coco.py',
'configs/nas_fcos/nas_fcos_nashead_r50_caffe_fpn_gn-head_4x4_1x_coco.py',
'configs/nas_fpn/retinanet_r50_nasfpn_crop640_50e_coco.py',
'configs/paa/paa_r50_fpn_1x_coco.py',
'configs/pafpn/faster_rcnn_r50_pafpn_1x_coco.py',
'configs/pisa/pisa_mask_rcnn_r50_fpn_1x_coco.py',
'configs/point_rend/point_rend_r50_caffe_fpn_mstrain_1x_coco.py',
'configs/regnet/mask_rcnn_regnetx-3GF_fpn_1x_coco.py',
'configs/reppoints/reppoints_moment_r50_fpn_gn-neck+head_1x_coco.py',
'configs/res2net/faster_rcnn_r2_101_fpn_2x_coco.py',
'configs/resnest/'
'mask_rcnn_s50_fpn_syncbn-backbone+head_mstrain_1x_coco.py',
'configs/retinanet/retinanet_r50_caffe_fpn_1x_coco.py',
'configs/rpn/rpn_r50_fpn_1x_coco.py',
'configs/sabl/sabl_retinanet_r50_fpn_1x_coco.py',
'configs/ssd/ssd300_coco.py',
'configs/tridentnet/tridentnet_r50_caffe_1x_coco.py',
'configs/vfnet/vfnet_r50_fpn_1x_coco.py',
'configs/yolact/yolact_r50_1x8_coco.py',
'configs/yolo/yolov3_d53_320_273e_coco.py'
]
def main():
args = parse_args()
benchmark_type = []
if args.basic_arch:
benchmark_type += basic_arch_root
if args.datasets:
benchmark_type += datasets_root
if args.data_pipeline:
benchmark_type += data_pipeline_root
if args.nn_module:
benchmark_type += nn_module_root
special_model = args.options
if special_model is not None:
benchmark_type += special_model
config_dpath = 'configs/'
benchmark_configs = []
for cfg_root in benchmark_type:
cfg_dir = osp.join(config_dpath, cfg_root)
configs = os.scandir(cfg_dir)
for cfg in configs:
config_path = osp.join(cfg_dir, cfg.name)
if (config_path in benchmark_pool
and config_path not in benchmark_configs):
benchmark_configs.append(config_path)
print(f'Totally found {len(benchmark_configs)} configs to benchmark')
config_dicts = dict(models=benchmark_configs)
mmcv.dump(config_dicts, 'regression_test_configs.json')
if __name__ == '__main__':
main()
PyTorch/NLP/Conformer-main/mmdetection/.dev_scripts/convert_benchmark_script.py 0 → 100644
View file @ a4372e17
import argparse
import os
import os.path as osp
import mmcv
def parse_args():
parser = argparse.ArgumentParser(
description='Convert benchmark model json to script')
parser.add_argument(
'json_path', type=str, help='json path output by benchmark_filter')
parser.add_argument('partition', type=str, help='slurm partition name')
parser.add_argument(
'--run', action='store_true', help='run script directly')
parser.add_argument(
'--out', type=str, help='path to save model benchmark script')
args = parser.parse_args()
return args
def main():
args = parse_args()
if args.out:
out_suffix = args.out.split('.')[-1]
assert args.out.endswith('.sh'), \
f'Expected out file path suffix is .sh, but get .{out_suffix}'
assert args.out or args.run, \
('Please specify at least one operation (save/run/ the '
'script) with the argument "--out" or "--run"')
json_data = mmcv.load(args.json_path)
model_cfgs = json_data['models']
partition = args.partition # cluster name
root_name = './tools'
train_script_name = osp.join(root_name, 'slurm_train.sh')
# stdout is no output
stdout_cfg = '>/dev/null'
commands = []
for i, cfg in enumerate(model_cfgs):
# print cfg name
echo_info = 'echo \'' + cfg + '\' &'
commands.append(echo_info)
commands.append('\n')
fname, _ = osp.splitext(osp.basename(cfg))
out_fname = osp.join(root_name, fname)
# default setting
command_info = 'GPUS=8 GPUS_PER_NODE=8 CPUS_PER_TASK=2 ' \
+ train_script_name + ' '
command_info += partition + ' '
command_info += fname + ' '
command_info += cfg + ' '
command_info += out_fname + ' '
command_info += stdout_cfg + ' &'
commands.append(command_info)
if i < len(model_cfgs):
commands.append('\n')
command_str = ''.join(commands)
if args.out:
with open(args.out, 'w') as f:
f.write(command_str)
if args.run:
os.system(command_str)
if __name__ == '__main__':
main()
PyTorch/NLP/Conformer-main/mmdetection/.dev_scripts/gather_models.py 0 → 100644
View file @ a4372e17
import argparse
import glob
import json
import os.path as osp
import shutil
import subprocess
import mmcv
import torch
def process_checkpoint(in_file, out_file):
checkpoint = torch.load(in_file, map_location='cpu')
# remove optimizer for smaller file size
if 'optimizer' in checkpoint:
del checkpoint['optimizer']
# if it is necessary to remove some sensitive data in checkpoint['meta'],
# add the code here.
torch.save(checkpoint, out_file)
sha = subprocess.check_output(['sha256sum', out_file]).decode()
final_file = out_file.rstrip('.pth') + '-{}.pth'.format(sha[:8])
subprocess.Popen(['mv', out_file, final_file])
return final_file
def get_final_epoch(config):
cfg = mmcv.Config.fromfile('./configs/' + config)
return cfg.total_epochs
def get_final_results(log_json_path, epoch, results_lut):
result_dict = dict()
with open(log_json_path, 'r') as f:
for line in f.readlines():
log_line = json.loads(line)
if 'mode' not in log_line.keys():
continue
if log_line['mode'] == 'train' and log_line['epoch'] == epoch:
result_dict['memory'] = log_line['memory']
if log_line['mode'] == 'val' and log_line['epoch'] == epoch:
result_dict.update({
key: log_line[key]
for key in results_lut if key in log_line
})
return result_dict
def parse_args():
parser = argparse.ArgumentParser(description='Gather benchmarked models')
parser.add_argument(
'root',
type=str,
help='root path of benchmarked models to be gathered')
parser.add_argument(
'out', type=str, help='output path of gathered models to be stored')
args = parser.parse_args()
return args
def main():
args = parse_args()
models_root = args.root
models_out = args.out
mmcv.mkdir_or_exist(models_out)
# find all models in the root directory to be gathered
raw_configs = list(mmcv.scandir('./configs', '.py', recursive=True))
# filter configs that is not trained in the experiments dir
used_configs = []
for raw_config in raw_configs:
if osp.exists(osp.join(models_root, raw_config)):
used_configs.append(raw_config)
print(f'Find {len(used_configs)} models to be gathered')
# find final_ckpt and log file for trained each config
# and parse the best performance
model_infos = []
for used_config in used_configs:
exp_dir = osp.join(models_root, used_config)
# check whether the exps is finished
final_epoch = get_final_epoch(used_config)
final_model = 'epoch_{}.pth'.format(final_epoch)
model_path = osp.join(exp_dir, final_model)
# skip if the model is still training
if not osp.exists(model_path):
continue
# get the latest logs
log_json_path = list(
sorted(glob.glob(osp.join(exp_dir, '*.log.json'))))[-1]
log_txt_path = list(sorted(glob.glob(osp.join(exp_dir, '*.log'))))[-1]
cfg = mmcv.Config.fromfile('./configs/' + used_config)
results_lut = cfg.evaluation.metric
if not isinstance(results_lut, list):
results_lut = [results_lut]
# case when using VOC, the evaluation key is only 'mAP'
results_lut = [key + '_mAP' for key in results_lut if 'mAP' not in key]
model_performance = get_final_results(log_json_path, final_epoch,
results_lut)
if model_performance is None:
continue
model_time = osp.split(log_txt_path)[-1].split('.')[0]
model_infos.append(
dict(
config=used_config,
results=model_performance,
epochs=final_epoch,
model_time=model_time,
log_json_path=osp.split(log_json_path)[-1]))
# publish model for each checkpoint
publish_model_infos = []
for model in model_infos:
model_publish_dir = osp.join(models_out, model['config'].rstrip('.py'))
mmcv.mkdir_or_exist(model_publish_dir)
model_name = osp.split(model['config'])[-1].split('.')[0]
model_name += '_' + model['model_time']
publish_model_path = osp.join(model_publish_dir, model_name)
trained_model_path = osp.join(models_root, model['config'],
'epoch_{}.pth'.format(model['epochs']))
# convert model
final_model_path = process_checkpoint(trained_model_path,
publish_model_path)
# copy log
shutil.copy(
osp.join(models_root, model['config'], model['log_json_path']),
osp.join(model_publish_dir, f'{model_name}.log.json'))
shutil.copy(
osp.join(models_root, model['config'],
model['log_json_path'].rstrip('.json')),
osp.join(model_publish_dir, f'{model_name}.log'))
# copy config to guarantee reproducibility
config_path = model['config']
config_path = osp.join(
'configs',
config_path) if 'configs' not in config_path else config_path
target_cconfig_path = osp.split(config_path)[-1]
shutil.copy(config_path,
osp.join(model_publish_dir, target_cconfig_path))
model['model_path'] = final_model_path
publish_model_infos.append(model)
models = dict(models=publish_model_infos)
print(f'Totally gathered {len(publish_model_infos)} models')
mmcv.dump(models, osp.join(models_out, 'model_info.json'))
if __name__ == '__main__':
main()
PyTorch/NLP/Conformer-main/mmdetection/.dev_scripts/linter.sh 0 → 100644
View file @ a4372e17
yapf -r -i mmdet/ configs/ tests/ tools/
isort -rc mmdet/ configs/ tests/ tools/
flake8 .
PyTorch/NLP/Conformer-main/mmdetection/.github/CODE_OF_CONDUCT.md 0 → 100644
View file @ a4372e17
# Contributor Covenant Code of Conduct
## Our Pledge
In the interest of fostering an open and welcoming environment, we as
contributors and maintainers pledge to making participation in our project and
our community a harassment-free experience for everyone, regardless of age, body
size, disability, ethnicity, sex characteristics, gender identity and expression,
level of experience, education, socio-economic status, nationality, personal
appearance, race, religion, or sexual identity and orientation.
## Our Standards
Examples of behavior that contributes to creating a positive environment
include:
* Using welcoming and inclusive language
* Being respectful of differing viewpoints and experiences
* Gracefully accepting constructive criticism
* Focusing on what is best for the community
* Showing empathy towards other community members
Examples of unacceptable behavior by participants include:
* The use of sexualized language or imagery and unwelcome sexual attention or
advances
* Trolling, insulting/derogatory comments, and personal or political attacks
* Public or private harassment
* Publishing others' private information, such as a physical or electronic
address, without explicit permission
* Other conduct which could reasonably be considered inappropriate in a
professional setting
## Our Responsibilities
Project maintainers are responsible for clarifying the standards of acceptable
behavior and are expected to take appropriate and fair corrective action in
response to any instances of unacceptable behavior.
Project maintainers have the right and responsibility to remove, edit, or
reject comments, commits, code, wiki edits, issues, and other contributions
that are not aligned to this Code of Conduct, or to ban temporarily or
permanently any contributor for other behaviors that they deem inappropriate,
threatening, offensive, or harmful.
## Scope
This Code of Conduct applies both within project spaces and in public spaces
when an individual is representing the project or its community. Examples of
representing a project or community include using an official project e-mail
address, posting via an official social media account, or acting as an appointed
representative at an online or offline event. Representation of a project may be
further defined and clarified by project maintainers.
## Enforcement
Instances of abusive, harassing, or otherwise unacceptable behavior may be
reported by contacting the project team at chenkaidev@gmail.com. All
complaints will be reviewed and investigated and will result in a response that
is deemed necessary and appropriate to the circumstances. The project team is
obligated to maintain confidentiality with regard to the reporter of an incident.
Further details of specific enforcement policies may be posted separately.
Project maintainers who do not follow or enforce the Code of Conduct in good
faith may face temporary or permanent repercussions as determined by other
members of the project's leadership.
## Attribution
This Code of Conduct is adapted from the [Contributor Covenant][homepage], version 1.4,
available at https://www.contributor-covenant.org/version/1/4/code-of-conduct.html
[homepage]: https://www.contributor-covenant.org
For answers to common questions about this code of conduct, see
https://www.contributor-covenant.org/faq
PyTorch/NLP/Conformer-main/mmdetection/.github/CONTRIBUTING.md 0 → 100644
View file @ a4372e17
# Contributing to mmdetection
All kinds of contributions are welcome, including but not limited to the following.
- Fixes (typo, bugs)
- New features and components
## Workflow
1. fork and pull the latest mmdetection
2. checkout a new branch (do not use master branch for PRs)
3. commit your changes
4. create a PR
Note
- If you plan to add some new features that involve large changes, it is encouraged to open an issue for discussion first.
- If you are the author of some papers and would like to include your method to mmdetection, please let us know (open an issue or contact the maintainers). We will much appreciate your contribution.
- For new features and new modules, unit tests are required to improve the code's robustness.
## Code style
### Python
We adopt [PEP8](https://www.python.org/dev/peps/pep-0008/) as the preferred code style.
We use the following tools for linting and formatting:
- [flake8](http://flake8.pycqa.org/en/latest/): linter
- [yapf](https://github.com/google/yapf): formatter
- [isort](https://github.com/timothycrosley/isort): sort imports
Style configurations of yapf and isort can be found in [setup.cfg](../setup.cfg).
We use [pre-commit hook](https://pre-commit.com/) that checks and formats for `flake8`, `yapf`, `isort`, `trailing whitespaces`,
fixes `end-of-files`, sorts `requirments.txt` automatically on every commit.
The config for a pre-commit hook is stored in [.pre-commit-config](../.pre-commit-config.yaml).
After you clone the repository, you will need to install initialize pre-commit hook.
```shell
pip install -U pre-commit
```
From the repository folder
```shell
pre-commit install
```
If you are facing issue when installing markdown lint, you may install ruby for markdown lint by following
```shell
# install rvm
curl -L https://get.rvm.io | bash -s -- --autolibs=read-fail
# set up environment
# Note that you might need to edit ~/.bashrc, ~/.bash_profile, and ~/.profile
source ~/.profile
rvm autolibs disable
# install ruby
rvm install 2.7.1
```
After this on every commit check code linters and formatter will be enforced.
>Before you create a PR, make sure that your code lints and is formatted by yapf.
### C++ and CUDA
We follow the [Google C++ Style Guide](https://google.github.io/styleguide/cppguide.html).
PyTorch/NLP/Conformer-main/mmdetection/.github/ISSUE_TEMPLATE/config.yml 0 → 100644
View file @ a4372e17
blank_issues_enabled: false
contact_links:
- name: Common Issues
url: https://mmdetection.readthedocs.io/en/latest/faq.html
about: Check if your issue already has solutions
- name: MMDetection Documentation
url: https://mmdetection.readthedocs.io/en/latest/
about: Check if your question is answered in docs
PyTorch/NLP/Conformer-main/mmdetection/.github/ISSUE_TEMPLATE/error-report.md 0 → 100644
View file @ a4372e17
---
name: Error report
about: Create a report to help us improve
title: ''
labels: ''
assignees: ''
---
Thanks for your error report and we appreciate it a lot.
**Checklist**
1. I have searched related issues but cannot get the expected help.
2. I have read the [FAQ documentation](https://mmdetection.readthedocs.io/en/latest/faq.html) but cannot get the expected help.
3. The bug has not been fixed in the latest version.
**Describe the bug**
A clear and concise description of what the bug is.
**Reproduction**
1. What command or script did you run?
```none
A placeholder for the command.
```
2. Did you make any modifications on the code or config? Did you understand what you have modified?
3. What dataset did you use?
**Environment**
1. Please run `python mmdet/utils/collect_env.py` to collect necessary environment information and paste it here.
2. You may add addition that may be helpful for locating the problem, such as
- How you installed PyTorch [e.g., pip, conda, source]
- Other environment variables that may be related (such as `$PATH`, `$LD_LIBRARY_PATH`, `$PYTHONPATH`, etc.)
**Error traceback**
If applicable, paste the error trackback here.
```none
A placeholder for trackback.
```
**Bug fix**
If you have already identified the reason, you can provide the information here. If you are willing to create a PR to fix it, please also leave a comment here and that would be much appreciated!
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment