Merge branch 'hepj_work' into 'main'

增加conformer代码

See merge request dcutoolkit/deeplearing/dlexamples_new!7

PyTorch/NLP/Conformer-main/mmdetection/mmdet/models/losses/gaussian_focal_loss.py

+import torch.nn as nn
+
+from ..builder import LOSSES
+from .utils import weighted_loss
+
+
+@weighted_loss
+def gaussian_focal_loss(pred, gaussian_target, alpha=2.0, gamma=4.0):
+    """`Focal Loss <https://arxiv.org/abs/1708.02002>`_ for targets in gaussian
+    distribution.
+
+    Args:
+        pred (torch.Tensor): The prediction.
+        gaussian_target (torch.Tensor): The learning target of the prediction
+            in gaussian distribution.
+        alpha (float, optional): A balanced form for Focal Loss.
+            Defaults to 2.0.
+        gamma (float, optional): The gamma for calculating the modulating
+            factor. Defaults to 4.0.
+    """
+    eps = 1e-12
+    pos_weights = gaussian_target.eq(1)
+    neg_weights = (1 - gaussian_target).pow(gamma)
+    pos_loss = -(pred + eps).log() * (1 - pred).pow(alpha) * pos_weights
+    neg_loss = -(1 - pred + eps).log() * pred.pow(alpha) * neg_weights
+    return pos_loss + neg_loss
+
+
+@LOSSES.register_module()
+class GaussianFocalLoss(nn.Module):
+    """GaussianFocalLoss is a variant of focal loss.
+
+    More details can be found in the `paper
+    <https://arxiv.org/abs/1808.01244>`_
+    Code is modified from `kp_utils.py
+    <https://github.com/princeton-vl/CornerNet/blob/master/models/py_utils/kp_utils.py#L152>`_  # noqa: E501
+    Please notice that the target in GaussianFocalLoss is a gaussian heatmap,
+    not 0/1 binary target.
+
+    Args:
+        alpha (float): Power of prediction.
+        gamma (float): Power of target for negtive samples.
+        reduction (str): Options are "none", "mean" and "sum".
+        loss_weight (float): Loss weight of current loss.
+    """
+
+    def __init__(self,
+                 alpha=2.0,
+                 gamma=4.0,
+                 reduction='mean',
+                 loss_weight=1.0):
+        super(GaussianFocalLoss, self).__init__()
+        self.alpha = alpha
+        self.gamma = gamma
+        self.reduction = reduction
+        self.loss_weight = loss_weight
+
+    def forward(self,
+                pred,
+                target,
+                weight=None,
+                avg_factor=None,
+                reduction_override=None):
+        """Forward function.
+
+        Args:
+            pred (torch.Tensor): The prediction.
+            target (torch.Tensor): The learning target of the prediction
+                in gaussian distribution.
+            weight (torch.Tensor, optional): The weight of loss for each
+                prediction. Defaults to None.
+            avg_factor (int, optional): Average factor that is used to average
+                the loss. Defaults to None.
+            reduction_override (str, optional): The reduction method used to
+                override the original reduction method of the loss.
+                Defaults to None.
+        """
+        assert reduction_override in (None, 'none', 'mean', 'sum')
+        reduction = (
+            reduction_override if reduction_override else self.reduction)
+        loss_reg = self.loss_weight * gaussian_focal_loss(
+            pred,
+            target,
+            weight,
+            alpha=self.alpha,
+            gamma=self.gamma,
+            reduction=reduction,
+            avg_factor=avg_factor)
+        return loss_reg

PyTorch/NLP/Conformer-main/mmdetection/mmdet/models/losses/gfocal_loss.py

+import torch.nn as nn
+import torch.nn.functional as F
+
+from ..builder import LOSSES
+from .utils import weighted_loss
+
+
+@weighted_loss
+def quality_focal_loss(pred, target, beta=2.0):
+    r"""Quality Focal Loss (QFL) is from `Generalized Focal Loss: Learning
+    Qualified and Distributed Bounding Boxes for Dense Object Detection
+    <https://arxiv.org/abs/2006.04388>`_.
+
+    Args:
+        pred (torch.Tensor): Predicted joint representation of classification
+            and quality (IoU) estimation with shape (N, C), C is the number of
+            classes.
+        target (tuple([torch.Tensor])): Target category label with shape (N,)
+            and target quality label with shape (N,).
+        beta (float): The beta parameter for calculating the modulating factor.
+            Defaults to 2.0.
+
+    Returns:
+        torch.Tensor: Loss tensor with shape (N,).
+    """
+    assert len(target) == 2, """target for QFL must be a tuple of two elements,
+        including category label and quality label, respectively"""
+    # label denotes the category id, score denotes the quality score
+    label, score = target
+
+    # negatives are supervised by 0 quality score
+    pred_sigmoid = pred.sigmoid()
+    scale_factor = pred_sigmoid
+    zerolabel = scale_factor.new_zeros(pred.shape)
+    loss = F.binary_cross_entropy_with_logits(
+        pred, zerolabel, reduction='none') * scale_factor.pow(beta)
+
+    # FG cat_id: [0, num_classes -1], BG cat_id: num_classes
+    bg_class_ind = pred.size(1)
+    pos = ((label >= 0) & (label < bg_class_ind)).nonzero().squeeze(1)
+    pos_label = label[pos].long()
+    # positives are supervised by bbox quality (IoU) score
+    scale_factor = score[pos] - pred_sigmoid[pos, pos_label]
+    loss[pos, pos_label] = F.binary_cross_entropy_with_logits(
+        pred[pos, pos_label], score[pos],
+        reduction='none') * scale_factor.abs().pow(beta)
+
+    loss = loss.sum(dim=1, keepdim=False)
+    return loss
+
+
+@weighted_loss
+def distribution_focal_loss(pred, label):
+    r"""Distribution Focal Loss (DFL) is from `Generalized Focal Loss: Learning
+    Qualified and Distributed Bounding Boxes for Dense Object Detection
+    <https://arxiv.org/abs/2006.04388>`_.
+
+    Args:
+        pred (torch.Tensor): Predicted general distribution of bounding boxes
+            (before softmax) with shape (N, n+1), n is the max value of the
+            integral set `{0, ..., n}` in paper.
+        label (torch.Tensor): Target distance label for bounding boxes with
+            shape (N,).
+
+    Returns:
+        torch.Tensor: Loss tensor with shape (N,).
+    """
+    dis_left = label.long()
+    dis_right = dis_left + 1
+    weight_left = dis_right.float() - label
+    weight_right = label - dis_left.float()
+    loss = F.cross_entropy(pred, dis_left, reduction='none') * weight_left \
+        + F.cross_entropy(pred, dis_right, reduction='none') * weight_right
+    return loss
+
+
+@LOSSES.register_module()
+class QualityFocalLoss(nn.Module):
+    r"""Quality Focal Loss (QFL) is a variant of `Generalized Focal Loss:
+    Learning Qualified and Distributed Bounding Boxes for Dense Object
+    Detection <https://arxiv.org/abs/2006.04388>`_.
+
+    Args:
+        use_sigmoid (bool): Whether sigmoid operation is conducted in QFL.
+            Defaults to True.
+        beta (float): The beta parameter for calculating the modulating factor.
+            Defaults to 2.0.
+        reduction (str): Options are "none", "mean" and "sum".
+        loss_weight (float): Loss weight of current loss.
+    """
+
+    def __init__(self,
+                 use_sigmoid=True,
+                 beta=2.0,
+                 reduction='mean',
+                 loss_weight=1.0):
+        super(QualityFocalLoss, self).__init__()
+        assert use_sigmoid is True, 'Only sigmoid in QFL supported now.'
+        self.use_sigmoid = use_sigmoid
+        self.beta = beta
+        self.reduction = reduction
+        self.loss_weight = loss_weight
+
+    def forward(self,
+                pred,
+                target,
+                weight=None,
+                avg_factor=None,
+                reduction_override=None):
+        """Forward function.
+
+        Args:
+            pred (torch.Tensor): Predicted joint representation of
+                classification and quality (IoU) estimation with shape (N, C),
+                C is the number of classes.
+            target (tuple([torch.Tensor])): Target category label with shape
+                (N,) and target quality label with shape (N,).
+            weight (torch.Tensor, optional): The weight of loss for each
+                prediction. Defaults to None.
+            avg_factor (int, optional): Average factor that is used to average
+                the loss. Defaults to None.
+            reduction_override (str, optional): The reduction method used to
+                override the original reduction method of the loss.
+                Defaults to None.
+        """
+        assert reduction_override in (None, 'none', 'mean', 'sum')
+        reduction = (
+            reduction_override if reduction_override else self.reduction)
+        if self.use_sigmoid:
+            loss_cls = self.loss_weight * quality_focal_loss(
+                pred,
+                target,
+                weight,
+                beta=self.beta,
+                reduction=reduction,
+                avg_factor=avg_factor)
+        else:
+            raise NotImplementedError
+        return loss_cls
+
+
+@LOSSES.register_module()
+class DistributionFocalLoss(nn.Module):
+    r"""Distribution Focal Loss (DFL) is a variant of `Generalized Focal Loss:
+    Learning Qualified and Distributed Bounding Boxes for Dense Object
+    Detection <https://arxiv.org/abs/2006.04388>`_.
+
+    Args:
+        reduction (str): Options are `'none'`, `'mean'` and `'sum'`.
+        loss_weight (float): Loss weight of current loss.
+    """
+
+    def __init__(self, reduction='mean', loss_weight=1.0):
+        super(DistributionFocalLoss, self).__init__()
+        self.reduction = reduction
+        self.loss_weight = loss_weight
+
+    def forward(self,
+                pred,
+                target,
+                weight=None,
+                avg_factor=None,
+                reduction_override=None):
+        """Forward function.
+
+        Args:
+            pred (torch.Tensor): Predicted general distribution of bounding
+                boxes (before softmax) with shape (N, n+1), n is the max value
+                of the integral set `{0, ..., n}` in paper.
+            target (torch.Tensor): Target distance label for bounding boxes
+                with shape (N,).
+            weight (torch.Tensor, optional): The weight of loss for each
+                prediction. Defaults to None.
+            avg_factor (int, optional): Average factor that is used to average
+                the loss. Defaults to None.
+            reduction_override (str, optional): The reduction method used to
+                override the original reduction method of the loss.
+                Defaults to None.
+        """
+        assert reduction_override in (None, 'none', 'mean', 'sum')
+        reduction = (
+            reduction_override if reduction_override else self.reduction)
+        loss_cls = self.loss_weight * distribution_focal_loss(
+            pred, target, weight, reduction=reduction, avg_factor=avg_factor)
+        return loss_cls

PyTorch/NLP/Conformer-main/mmdetection/mmdet/models/losses/ghm_loss.py

+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+
+from ..builder import LOSSES
+
+
+def _expand_onehot_labels(labels, label_weights, label_channels):
+    bin_labels = labels.new_full((labels.size(0), label_channels), 0)
+    inds = torch.nonzero(
+        (labels >= 0) & (labels < label_channels), as_tuple=False).squeeze()
+    if inds.numel() > 0:
+        bin_labels[inds, labels[inds]] = 1
+    bin_label_weights = label_weights.view(-1, 1).expand(
+        label_weights.size(0), label_channels)
+    return bin_labels, bin_label_weights
+
+
+# TODO: code refactoring to make it consistent with other losses
+@LOSSES.register_module()
+class GHMC(nn.Module):
+    """GHM Classification Loss.
+
+    Details of the theorem can be viewed in the paper
+    `Gradient Harmonized Single-stage Detector
+    <https://arxiv.org/abs/1811.05181>`_.
+
+    Args:
+        bins (int): Number of the unit regions for distribution calculation.
+        momentum (float): The parameter for moving average.
+        use_sigmoid (bool): Can only be true for BCE based loss now.
+        loss_weight (float): The weight of the total GHM-C loss.
+    """
+
+    def __init__(self, bins=10, momentum=0, use_sigmoid=True, loss_weight=1.0):
+        super(GHMC, self).__init__()
+        self.bins = bins
+        self.momentum = momentum
+        edges = torch.arange(bins + 1).float() / bins
+        self.register_buffer('edges', edges)
+        self.edges[-1] += 1e-6
+        if momentum > 0:
+            acc_sum = torch.zeros(bins)
+            self.register_buffer('acc_sum', acc_sum)
+        self.use_sigmoid = use_sigmoid
+        if not self.use_sigmoid:
+            raise NotImplementedError
+        self.loss_weight = loss_weight
+
+    def forward(self, pred, target, label_weight, *args, **kwargs):
+        """Calculate the GHM-C loss.
+
+        Args:
+            pred (float tensor of size [batch_num, class_num]):
+                The direct prediction of classification fc layer.
+            target (float tensor of size [batch_num, class_num]):
+                Binary class target for each sample.
+            label_weight (float tensor of size [batch_num, class_num]):
+                the value is 1 if the sample is valid and 0 if ignored.
+        Returns:
+            The gradient harmonized loss.
+        """
+        # the target should be binary class label
+        if pred.dim() != target.dim():
+            target, label_weight = _expand_onehot_labels(
+                target, label_weight, pred.size(-1))
+        target, label_weight = target.float(), label_weight.float()
+        edges = self.edges
+        mmt = self.momentum
+        weights = torch.zeros_like(pred)
+
+        # gradient length
+        g = torch.abs(pred.sigmoid().detach() - target)
+
+        valid = label_weight > 0
+        tot = max(valid.float().sum().item(), 1.0)
+        n = 0  # n valid bins
+        for i in range(self.bins):
+            inds = (g >= edges[i]) & (g < edges[i + 1]) & valid
+            num_in_bin = inds.sum().item()
+            if num_in_bin > 0:
+                if mmt > 0:
+                    self.acc_sum[i] = mmt * self.acc_sum[i] \
+                        + (1 - mmt) * num_in_bin
+                    weights[inds] = tot / self.acc_sum[i]
+                else:
+                    weights[inds] = tot / num_in_bin
+                n += 1
+        if n > 0:
+            weights = weights / n
+
+        loss = F.binary_cross_entropy_with_logits(
+            pred, target, weights, reduction='sum') / tot
+        return loss * self.loss_weight
+
+
+# TODO: code refactoring to make it consistent with other losses
+@LOSSES.register_module()
+class GHMR(nn.Module):
+    """GHM Regression Loss.
+
+    Details of the theorem can be viewed in the paper
+    `Gradient Harmonized Single-stage Detector
+    <https://arxiv.org/abs/1811.05181>`_.
+
+    Args:
+        mu (float): The parameter for the Authentic Smooth L1 loss.
+        bins (int): Number of the unit regions for distribution calculation.
+        momentum (float): The parameter for moving average.
+        loss_weight (float): The weight of the total GHM-R loss.
+    """
+
+    def __init__(self, mu=0.02, bins=10, momentum=0, loss_weight=1.0):
+        super(GHMR, self).__init__()
+        self.mu = mu
+        self.bins = bins
+        edges = torch.arange(bins + 1).float() / bins
+        self.register_buffer('edges', edges)
+        self.edges[-1] = 1e3
+        self.momentum = momentum
+        if momentum > 0:
+            acc_sum = torch.zeros(bins)
+            self.register_buffer('acc_sum', acc_sum)
+        self.loss_weight = loss_weight
+
+    # TODO: support reduction parameter
+    def forward(self, pred, target, label_weight, avg_factor=None):
+        """Calculate the GHM-R loss.
+
+        Args:
+            pred (float tensor of size [batch_num, 4 (* class_num)]):
+                The prediction of box regression layer. Channel number can be 4
+                or 4 * class_num depending on whether it is class-agnostic.
+            target (float tensor of size [batch_num, 4 (* class_num)]):
+                The target regression values with the same size of pred.
+            label_weight (float tensor of size [batch_num, 4 (* class_num)]):
+                The weight of each sample, 0 if ignored.
+        Returns:
+            The gradient harmonized loss.
+        """
+        mu = self.mu
+        edges = self.edges
+        mmt = self.momentum
+
+        # ASL1 loss
+        diff = pred - target
+        loss = torch.sqrt(diff * diff + mu * mu) - mu
+
+        # gradient length
+        g = torch.abs(diff / torch.sqrt(mu * mu + diff * diff)).detach()
+        weights = torch.zeros_like(g)
+
+        valid = label_weight > 0
+        tot = max(label_weight.float().sum().item(), 1.0)
+        n = 0  # n: valid bins
+        for i in range(self.bins):
+            inds = (g >= edges[i]) & (g < edges[i + 1]) & valid
+            num_in_bin = inds.sum().item()
+            if num_in_bin > 0:
+                n += 1
+                if mmt > 0:
+                    self.acc_sum[i] = mmt * self.acc_sum[i] \
+                        + (1 - mmt) * num_in_bin
+                    weights[inds] = tot / self.acc_sum[i]
+                else:
+                    weights[inds] = tot / num_in_bin
+        if n > 0:
+            weights /= n
+
+        loss = loss * weights
+        loss = loss.sum() / tot
+        return loss * self.loss_weight

PyTorch/NLP/Conformer-main/mmdetection/mmdet/models/losses/iou_loss.py

+import math
+
+import torch
+import torch.nn as nn
+
+from mmdet.core import bbox_overlaps
+from ..builder import LOSSES
+from .utils import weighted_loss
+
+
+@weighted_loss
+def iou_loss(pred, target, linear=False, eps=1e-6):
+    """IoU loss.
+
+    Computing the IoU loss between a set of predicted bboxes and target bboxes.
+    The loss is calculated as negative log of IoU.
+
+    Args:
+        pred (torch.Tensor): Predicted bboxes of format (x1, y1, x2, y2),
+            shape (n, 4).
+        target (torch.Tensor): Corresponding gt bboxes, shape (n, 4).
+        linear (bool, optional): If True, use linear scale of loss instead of
+            log scale. Default: False.
+        eps (float): Eps to avoid log(0).
+
+    Return:
+        torch.Tensor: Loss tensor.
+    """
+    ious = bbox_overlaps(pred, target, is_aligned=True).clamp(min=eps)
+    if linear:
+        loss = 1 - ious
+    else:
+        loss = -ious.log()
+    return loss
+
+
+@weighted_loss
+def bounded_iou_loss(pred, target, beta=0.2, eps=1e-3):
+    """BIoULoss.
+
+    This is an implementation of paper
+    `Improving Object Localization with Fitness NMS and Bounded IoU Loss.
+    <https://arxiv.org/abs/1711.00164>`_.
+
+    Args:
+        pred (torch.Tensor): Predicted bboxes.
+        target (torch.Tensor): Target bboxes.
+        beta (float): beta parameter in smoothl1.
+        eps (float): eps to avoid NaN.
+    """
+    pred_ctrx = (pred[:, 0] + pred[:, 2]) * 0.5
+    pred_ctry = (pred[:, 1] + pred[:, 3]) * 0.5
+    pred_w = pred[:, 2] - pred[:, 0]
+    pred_h = pred[:, 3] - pred[:, 1]
+    with torch.no_grad():
+        target_ctrx = (target[:, 0] + target[:, 2]) * 0.5
+        target_ctry = (target[:, 1] + target[:, 3]) * 0.5
+        target_w = target[:, 2] - target[:, 0]
+        target_h = target[:, 3] - target[:, 1]
+
+    dx = target_ctrx - pred_ctrx
+    dy = target_ctry - pred_ctry
+
+    loss_dx = 1 - torch.max(
+        (target_w - 2 * dx.abs()) /
+        (target_w + 2 * dx.abs() + eps), torch.zeros_like(dx))
+    loss_dy = 1 - torch.max(
+        (target_h - 2 * dy.abs()) /
+        (target_h + 2 * dy.abs() + eps), torch.zeros_like(dy))
+    loss_dw = 1 - torch.min(target_w / (pred_w + eps), pred_w /
+                            (target_w + eps))
+    loss_dh = 1 - torch.min(target_h / (pred_h + eps), pred_h /
+                            (target_h + eps))
+    loss_comb = torch.stack([loss_dx, loss_dy, loss_dw, loss_dh],
+                            dim=-1).view(loss_dx.size(0), -1)
+
+    loss = torch.where(loss_comb < beta, 0.5 * loss_comb * loss_comb / beta,
+                       loss_comb - 0.5 * beta)
+    return loss
+
+
+@weighted_loss
+def giou_loss(pred, target, eps=1e-7):
+    r"""`Generalized Intersection over Union: A Metric and A Loss for Bounding
+    Box Regression <https://arxiv.org/abs/1902.09630>`_.
+
+    Args:
+        pred (torch.Tensor): Predicted bboxes of format (x1, y1, x2, y2),
+            shape (n, 4).
+        target (torch.Tensor): Corresponding gt bboxes, shape (n, 4).
+        eps (float): Eps to avoid log(0).
+
+    Return:
+        Tensor: Loss tensor.
+    """
+    gious = bbox_overlaps(pred, target, mode='giou', is_aligned=True, eps=eps)
+    loss = 1 - gious
+    return loss
+
+
+@weighted_loss
+def diou_loss(pred, target, eps=1e-7):
+    r"""`Implementation of Distance-IoU Loss: Faster and Better
+    Learning for Bounding Box Regression, https://arxiv.org/abs/1911.08287`_.
+
+    Code is modified from https://github.com/Zzh-tju/DIoU.
+
+    Args:
+        pred (Tensor): Predicted bboxes of format (x1, y1, x2, y2),
+            shape (n, 4).
+        target (Tensor): Corresponding gt bboxes, shape (n, 4).
+        eps (float): Eps to avoid log(0).
+    Return:
+        Tensor: Loss tensor.
+    """
+    # overlap
+    lt = torch.max(pred[:, :2], target[:, :2])
+    rb = torch.min(pred[:, 2:], target[:, 2:])
+    wh = (rb - lt).clamp(min=0)
+    overlap = wh[:, 0] * wh[:, 1]
+
+    # union
+    ap = (pred[:, 2] - pred[:, 0]) * (pred[:, 3] - pred[:, 1])
+    ag = (target[:, 2] - target[:, 0]) * (target[:, 3] - target[:, 1])
+    union = ap + ag - overlap + eps
+
+    # IoU
+    ious = overlap / union
+
+    # enclose area
+    enclose_x1y1 = torch.min(pred[:, :2], target[:, :2])
+    enclose_x2y2 = torch.max(pred[:, 2:], target[:, 2:])
+    enclose_wh = (enclose_x2y2 - enclose_x1y1).clamp(min=0)
+
+    cw = enclose_wh[:, 0]
+    ch = enclose_wh[:, 1]
+
+    c2 = cw**2 + ch**2 + eps
+
+    b1_x1, b1_y1 = pred[:, 0], pred[:, 1]
+    b1_x2, b1_y2 = pred[:, 2], pred[:, 3]
+    b2_x1, b2_y1 = target[:, 0], target[:, 1]
+    b2_x2, b2_y2 = target[:, 2], target[:, 3]
+
+    left = ((b2_x1 + b2_x2) - (b1_x1 + b1_x2))**2 / 4
+    right = ((b2_y1 + b2_y2) - (b1_y1 + b1_y2))**2 / 4
+    rho2 = left + right
+
+    # DIoU
+    dious = ious - rho2 / c2
+    loss = 1 - dious
+    return loss
+
+
+@weighted_loss
+def ciou_loss(pred, target, eps=1e-7):
+    r"""`Implementation of paper `Enhancing Geometric Factors into
+    Model Learning and Inference for Object Detection and Instance
+    Segmentation <https://arxiv.org/abs/2005.03572>`_.
+
+    Code is modified from https://github.com/Zzh-tju/CIoU.
+
+    Args:
+        pred (Tensor): Predicted bboxes of format (x1, y1, x2, y2),
+            shape (n, 4).
+        target (Tensor): Corresponding gt bboxes, shape (n, 4).
+        eps (float): Eps to avoid log(0).
+    Return:
+        Tensor: Loss tensor.
+    """
+    # overlap
+    lt = torch.max(pred[:, :2], target[:, :2])
+    rb = torch.min(pred[:, 2:], target[:, 2:])
+    wh = (rb - lt).clamp(min=0)
+    overlap = wh[:, 0] * wh[:, 1]
+
+    # union
+    ap = (pred[:, 2] - pred[:, 0]) * (pred[:, 3] - pred[:, 1])
+    ag = (target[:, 2] - target[:, 0]) * (target[:, 3] - target[:, 1])
+    union = ap + ag - overlap + eps
+
+    # IoU
+    ious = overlap / union
+
+    # enclose area
+    enclose_x1y1 = torch.min(pred[:, :2], target[:, :2])
+    enclose_x2y2 = torch.max(pred[:, 2:], target[:, 2:])
+    enclose_wh = (enclose_x2y2 - enclose_x1y1).clamp(min=0)
+
+    cw = enclose_wh[:, 0]
+    ch = enclose_wh[:, 1]
+
+    c2 = cw**2 + ch**2 + eps
+
+    b1_x1, b1_y1 = pred[:, 0], pred[:, 1]
+    b1_x2, b1_y2 = pred[:, 2], pred[:, 3]
+    b2_x1, b2_y1 = target[:, 0], target[:, 1]
+    b2_x2, b2_y2 = target[:, 2], target[:, 3]
+
+    w1, h1 = b1_x2 - b1_x1, b1_y2 - b1_y1 + eps
+    w2, h2 = b2_x2 - b2_x1, b2_y2 - b2_y1 + eps
+
+    left = ((b2_x1 + b2_x2) - (b1_x1 + b1_x2))**2 / 4
+    right = ((b2_y1 + b2_y2) - (b1_y1 + b1_y2))**2 / 4
+    rho2 = left + right
+
+    factor = 4 / math.pi**2
+    v = factor * torch.pow(torch.atan(w2 / h2) - torch.atan(w1 / h1), 2)
+
+    # CIoU
+    cious = ious - (rho2 / c2 + v**2 / (1 - ious + v))
+    loss = 1 - cious
+    return loss
+
+
+@LOSSES.register_module()
+class IoULoss(nn.Module):
+    """IoULoss.
+
+    Computing the IoU loss between a set of predicted bboxes and target bboxes.
+
+    Args:
+        linear (bool): If True, use linear scale of loss instead of log scale.
+            Default: False.
+        eps (float): Eps to avoid log(0).
+        reduction (str): Options are "none", "mean" and "sum".
+        loss_weight (float): Weight of loss.
+    """
+
+    def __init__(self,
+                 linear=False,
+                 eps=1e-6,
+                 reduction='mean',
+                 loss_weight=1.0):
+        super(IoULoss, self).__init__()
+        self.linear = linear
+        self.eps = eps
+        self.reduction = reduction
+        self.loss_weight = loss_weight
+
+    def forward(self,
+                pred,
+                target,
+                weight=None,
+                avg_factor=None,
+                reduction_override=None,
+                **kwargs):
+        """Forward function.
+
+        Args:
+            pred (torch.Tensor): The prediction.
+            target (torch.Tensor): The learning target of the prediction.
+            weight (torch.Tensor, optional): The weight of loss for each
+                prediction. Defaults to None.
+            avg_factor (int, optional): Average factor that is used to average
+                the loss. Defaults to None.
+            reduction_override (str, optional): The reduction method used to
+                override the original reduction method of the loss.
+                Defaults to None. Options are "none", "mean" and "sum".
+        """
+        assert reduction_override in (None, 'none', 'mean', 'sum')
+        reduction = (
+            reduction_override if reduction_override else self.reduction)
+        if (weight is not None) and (not torch.any(weight > 0)) and (
+                reduction != 'none'):
+            return (pred * weight).sum()  # 0
+        if weight is not None and weight.dim() > 1:
+            # TODO: remove this in the future
+            # reduce the weight of shape (n, 4) to (n,) to match the
+            # iou_loss of shape (n,)
+            assert weight.shape == pred.shape
+            weight = weight.mean(-1)
+        loss = self.loss_weight * iou_loss(
+            pred,
+            target,
+            weight,
+            linear=self.linear,
+            eps=self.eps,
+            reduction=reduction,
+            avg_factor=avg_factor,
+            **kwargs)
+        return loss
+
+
+@LOSSES.register_module()
+class BoundedIoULoss(nn.Module):
+
+    def __init__(self, beta=0.2, eps=1e-3, reduction='mean', loss_weight=1.0):
+        super(BoundedIoULoss, self).__init__()
+        self.beta = beta
+        self.eps = eps
+        self.reduction = reduction
+        self.loss_weight = loss_weight
+
+    def forward(self,
+                pred,
+                target,
+                weight=None,
+                avg_factor=None,
+                reduction_override=None,
+                **kwargs):
+        if weight is not None and not torch.any(weight > 0):
+            return (pred * weight).sum()  # 0
+        assert reduction_override in (None, 'none', 'mean', 'sum')
+        reduction = (
+            reduction_override if reduction_override else self.reduction)
+        loss = self.loss_weight * bounded_iou_loss(
+            pred,
+            target,
+            weight,
+            beta=self.beta,
+            eps=self.eps,
+            reduction=reduction,
+            avg_factor=avg_factor,
+            **kwargs)
+        return loss
+
+
+@LOSSES.register_module()
+class GIoULoss(nn.Module):
+
+    def __init__(self, eps=1e-6, reduction='mean', loss_weight=1.0):
+        super(GIoULoss, self).__init__()
+        self.eps = eps
+        self.reduction = reduction
+        self.loss_weight = loss_weight
+
+    def forward(self,
+                pred,
+                target,
+                weight=None,
+                avg_factor=None,
+                reduction_override=None,
+                **kwargs):
+        if weight is not None and not torch.any(weight > 0):
+            return (pred * weight).sum()  # 0
+        assert reduction_override in (None, 'none', 'mean', 'sum')
+        reduction = (
+            reduction_override if reduction_override else self.reduction)
+        if weight is not None and weight.dim() > 1:
+            # TODO: remove this in the future
+            # reduce the weight of shape (n, 4) to (n,) to match the
+            # giou_loss of shape (n,)
+            assert weight.shape == pred.shape
+            weight = weight.mean(-1)
+        loss = self.loss_weight * giou_loss(
+            pred,
+            target,
+            weight,
+            eps=self.eps,
+            reduction=reduction,
+            avg_factor=avg_factor,
+            **kwargs)
+        return loss
+
+
+@LOSSES.register_module()
+class DIoULoss(nn.Module):
+
+    def __init__(self, eps=1e-6, reduction='mean', loss_weight=1.0):
+        super(DIoULoss, self).__init__()
+        self.eps = eps
+        self.reduction = reduction
+        self.loss_weight = loss_weight
+
+    def forward(self,
+                pred,
+                target,
+                weight=None,
+                avg_factor=None,
+                reduction_override=None,
+                **kwargs):
+        if weight is not None and not torch.any(weight > 0):
+            return (pred * weight).sum()  # 0
+        assert reduction_override in (None, 'none', 'mean', 'sum')
+        reduction = (
+            reduction_override if reduction_override else self.reduction)
+        if weight is not None and weight.dim() > 1:
+            # TODO: remove this in the future
+            # reduce the weight of shape (n, 4) to (n,) to match the
+            # giou_loss of shape (n,)
+            assert weight.shape == pred.shape
+            weight = weight.mean(-1)
+        loss = self.loss_weight * diou_loss(
+            pred,
+            target,
+            weight,
+            eps=self.eps,
+            reduction=reduction,
+            avg_factor=avg_factor,
+            **kwargs)
+        return loss
+
+
+@LOSSES.register_module()
+class CIoULoss(nn.Module):
+
+    def __init__(self, eps=1e-6, reduction='mean', loss_weight=1.0):
+        super(CIoULoss, self).__init__()
+        self.eps = eps
+        self.reduction = reduction
+        self.loss_weight = loss_weight
+
+    def forward(self,
+                pred,
+                target,
+                weight=None,
+                avg_factor=None,
+                reduction_override=None,
+                **kwargs):
+        if weight is not None and not torch.any(weight > 0):
+            return (pred * weight).sum()  # 0
+        assert reduction_override in (None, 'none', 'mean', 'sum')
+        reduction = (
+            reduction_override if reduction_override else self.reduction)
+        if weight is not None and weight.dim() > 1:
+            # TODO: remove this in the future
+            # reduce the weight of shape (n, 4) to (n,) to match the
+            # giou_loss of shape (n,)
+            assert weight.shape == pred.shape
+            weight = weight.mean(-1)
+        loss = self.loss_weight * ciou_loss(
+            pred,
+            target,
+            weight,
+            eps=self.eps,
+            reduction=reduction,
+            avg_factor=avg_factor,
+            **kwargs)
+        return loss

PyTorch/NLP/Conformer-main/mmdetection/mmdet/models/losses/mse_loss.py

+import torch.nn as nn
+import torch.nn.functional as F
+
+from ..builder import LOSSES
+from .utils import weighted_loss
+
+
+@weighted_loss
+def mse_loss(pred, target):
+    """Warpper of mse loss."""
+    return F.mse_loss(pred, target, reduction='none')
+
+
+@LOSSES.register_module()
+class MSELoss(nn.Module):
+    """MSELoss.
+
+    Args:
+        reduction (str, optional): The method that reduces the loss to a
+            scalar. Options are "none", "mean" and "sum".
+        loss_weight (float, optional): The weight of the loss. Defaults to 1.0
+    """
+
+    def __init__(self, reduction='mean', loss_weight=1.0):
+        super().__init__()
+        self.reduction = reduction
+        self.loss_weight = loss_weight
+
+    def forward(self, pred, target, weight=None, avg_factor=None):
+        """Forward function of loss.
+
+        Args:
+            pred (torch.Tensor): The prediction.
+            target (torch.Tensor): The learning target of the prediction.
+            weight (torch.Tensor, optional): Weight of the loss for each
+                prediction. Defaults to None.
+            avg_factor (int, optional): Average factor that is used to average
+                the loss. Defaults to None.
+
+        Returns:
+            torch.Tensor: The calculated loss
+        """
+        loss = self.loss_weight * mse_loss(
+            pred,
+            target,
+            weight,
+            reduction=self.reduction,
+            avg_factor=avg_factor)
+        return loss

PyTorch/NLP/Conformer-main/mmdetection/mmdet/models/losses/pisa_loss.py

+import torch
+
+from mmdet.core import bbox_overlaps
+
+
+def isr_p(cls_score,
+          bbox_pred,
+          bbox_targets,
+          rois,
+          sampling_results,
+          loss_cls,
+          bbox_coder,
+          k=2,
+          bias=0,
+          num_class=80):
+    """Importance-based Sample Reweighting (ISR_P), positive part.
+
+    Args:
+        cls_score (Tensor): Predicted classification scores.
+        bbox_pred (Tensor): Predicted bbox deltas.
+        bbox_targets (tuple[Tensor]): A tuple of bbox targets, the are
+            labels, label_weights, bbox_targets, bbox_weights, respectively.
+        rois (Tensor): Anchors (single_stage) in shape (n, 4) or RoIs
+            (two_stage) in shape (n, 5).
+        sampling_results (obj): Sampling results.
+        loss_cls (func): Classification loss func of the head.
+        bbox_coder (obj): BBox coder of the head.
+        k (float): Power of the non-linear mapping.
+        bias (float): Shift of the non-linear mapping.
+        num_class (int): Number of classes, default: 80.
+
+    Return:
+        tuple([Tensor]): labels, imp_based_label_weights, bbox_targets,
+            bbox_target_weights
+    """
+
+    labels, label_weights, bbox_targets, bbox_weights = bbox_targets
+    pos_label_inds = ((labels >= 0) &
+                      (labels < num_class)).nonzero().reshape(-1)
+    pos_labels = labels[pos_label_inds]
+
+    # if no positive samples, return the original targets
+    num_pos = float(pos_label_inds.size(0))
+    if num_pos == 0:
+        return labels, label_weights, bbox_targets, bbox_weights
+
+    # merge pos_assigned_gt_inds of per image to a single tensor
+    gts = list()
+    last_max_gt = 0
+    for i in range(len(sampling_results)):
+        gt_i = sampling_results[i].pos_assigned_gt_inds
+        gts.append(gt_i + last_max_gt)
+        if len(gt_i) != 0:
+            last_max_gt = gt_i.max() + 1
+    gts = torch.cat(gts)
+    assert len(gts) == num_pos
+
+    cls_score = cls_score.detach()
+    bbox_pred = bbox_pred.detach()
+
+    # For single stage detectors, rois here indicate anchors, in shape (N, 4)
+    # For two stage detectors, rois are in shape (N, 5)
+    if rois.size(-1) == 5:
+        pos_rois = rois[pos_label_inds][:, 1:]
+    else:
+        pos_rois = rois[pos_label_inds]
+
+    if bbox_pred.size(-1) > 4:
+        bbox_pred = bbox_pred.view(bbox_pred.size(0), -1, 4)
+        pos_delta_pred = bbox_pred[pos_label_inds, pos_labels].view(-1, 4)
+    else:
+        pos_delta_pred = bbox_pred[pos_label_inds].view(-1, 4)
+
+    # compute iou of the predicted bbox and the corresponding GT
+    pos_delta_target = bbox_targets[pos_label_inds].view(-1, 4)
+    pos_bbox_pred = bbox_coder.decode(pos_rois, pos_delta_pred)
+    target_bbox_pred = bbox_coder.decode(pos_rois, pos_delta_target)
+    ious = bbox_overlaps(pos_bbox_pred, target_bbox_pred, is_aligned=True)
+
+    pos_imp_weights = label_weights[pos_label_inds]
+    # Two steps to compute IoU-HLR. Samples are first sorted by IoU locally,
+    # then sorted again within the same-rank group
+    max_l_num = pos_labels.bincount().max()
+    for label in pos_labels.unique():
+        l_inds = (pos_labels == label).nonzero().view(-1)
+        l_gts = gts[l_inds]
+        for t in l_gts.unique():
+            t_inds = l_inds[l_gts == t]
+            t_ious = ious[t_inds]
+            _, t_iou_rank_idx = t_ious.sort(descending=True)
+            _, t_iou_rank = t_iou_rank_idx.sort()
+            ious[t_inds] += max_l_num - t_iou_rank.float()
+        l_ious = ious[l_inds]
+        _, l_iou_rank_idx = l_ious.sort(descending=True)
+        _, l_iou_rank = l_iou_rank_idx.sort()  # IoU-HLR
+        # linearly map HLR to label weights
+        pos_imp_weights[l_inds] *= (max_l_num - l_iou_rank.float()) / max_l_num
+
+    pos_imp_weights = (bias + pos_imp_weights * (1 - bias)).pow(k)
+
+    # normalize to make the new weighted loss value equal to the original loss
+    pos_loss_cls = loss_cls(
+        cls_score[pos_label_inds], pos_labels, reduction_override='none')
+    if pos_loss_cls.dim() > 1:
+        ori_pos_loss_cls = pos_loss_cls * label_weights[pos_label_inds][:,
+                                                                        None]
+        new_pos_loss_cls = pos_loss_cls * pos_imp_weights[:, None]
+    else:
+        ori_pos_loss_cls = pos_loss_cls * label_weights[pos_label_inds]
+        new_pos_loss_cls = pos_loss_cls * pos_imp_weights
+    pos_loss_cls_ratio = ori_pos_loss_cls.sum() / new_pos_loss_cls.sum()
+    pos_imp_weights = pos_imp_weights * pos_loss_cls_ratio
+    label_weights[pos_label_inds] = pos_imp_weights
+
+    bbox_targets = labels, label_weights, bbox_targets, bbox_weights
+    return bbox_targets
+
+
+def carl_loss(cls_score,
+              labels,
+              bbox_pred,
+              bbox_targets,
+              loss_bbox,
+              k=1,
+              bias=0.2,
+              avg_factor=None,
+              sigmoid=False,
+              num_class=80):
+    """Classification-Aware Regression Loss (CARL).
+
+    Args:
+        cls_score (Tensor): Predicted classification scores.
+        labels (Tensor): Targets of classification.
+        bbox_pred (Tensor): Predicted bbox deltas.
+        bbox_targets (Tensor): Target of bbox regression.
+        loss_bbox (func): Regression loss func of the head.
+        bbox_coder (obj): BBox coder of the head.
+        k (float): Power of the non-linear mapping.
+        bias (float): Shift of the non-linear mapping.
+        avg_factor (int): Average factor used in regression loss.
+        sigmoid (bool): Activation of the classification score.
+        num_class (int): Number of classes, default: 80.
+
+    Return:
+        dict: CARL loss dict.
+    """
+    pos_label_inds = ((labels >= 0) &
+                      (labels < num_class)).nonzero().reshape(-1)
+    if pos_label_inds.numel() == 0:
+        return dict(loss_carl=cls_score.sum()[None] * 0.)
+    pos_labels = labels[pos_label_inds]
+
+    # multiply pos_cls_score with the corresponding bbox weight
+    # and remain gradient
+    if sigmoid:
+        pos_cls_score = cls_score.sigmoid()[pos_label_inds, pos_labels]
+    else:
+        pos_cls_score = cls_score.softmax(-1)[pos_label_inds, pos_labels]
+    carl_loss_weights = (bias + (1 - bias) * pos_cls_score).pow(k)
+
+    # normalize carl_loss_weight to make its sum equal to num positive
+    num_pos = float(pos_cls_score.size(0))
+    weight_ratio = num_pos / carl_loss_weights.sum()
+    carl_loss_weights *= weight_ratio
+
+    if avg_factor is None:
+        avg_factor = bbox_targets.size(0)
+    # if is class agnostic, bbox pred is in shape (N, 4)
+    # otherwise, bbox pred is in shape (N, #classes, 4)
+    if bbox_pred.size(-1) > 4:
+        bbox_pred = bbox_pred.view(bbox_pred.size(0), -1, 4)
+        pos_bbox_preds = bbox_pred[pos_label_inds, pos_labels]
+    else:
+        pos_bbox_preds = bbox_pred[pos_label_inds]
+    ori_loss_reg = loss_bbox(
+        pos_bbox_preds,
+        bbox_targets[pos_label_inds],
+        reduction_override='none') / avg_factor
+    loss_carl = (ori_loss_reg * carl_loss_weights[:, None]).sum()
+    return dict(loss_carl=loss_carl[None])

PyTorch/NLP/Conformer-main/mmdetection/mmdet/models/losses/smooth_l1_loss.py

+import torch
+import torch.nn as nn
+
+from ..builder import LOSSES
+from .utils import weighted_loss
+
+
+@weighted_loss
+def smooth_l1_loss(pred, target, beta=1.0):
+    """Smooth L1 loss.
+
+    Args:
+        pred (torch.Tensor): The prediction.
+        target (torch.Tensor): The learning target of the prediction.
+        beta (float, optional): The threshold in the piecewise function.
+            Defaults to 1.0.
+
+    Returns:
+        torch.Tensor: Calculated loss
+    """
+    assert beta > 0
+    assert pred.size() == target.size() and target.numel() > 0
+    diff = torch.abs(pred - target)
+    loss = torch.where(diff < beta, 0.5 * diff * diff / beta,
+                       diff - 0.5 * beta)
+    return loss
+
+
+@weighted_loss
+def l1_loss(pred, target):
+    """L1 loss.
+
+    Args:
+        pred (torch.Tensor): The prediction.
+        target (torch.Tensor): The learning target of the prediction.
+
+    Returns:
+        torch.Tensor: Calculated loss
+    """
+    assert pred.size() == target.size() and target.numel() > 0
+    loss = torch.abs(pred - target)
+    return loss
+
+
+@LOSSES.register_module()
+class SmoothL1Loss(nn.Module):
+    """Smooth L1 loss.
+
+    Args:
+        beta (float, optional): The threshold in the piecewise function.
+            Defaults to 1.0.
+        reduction (str, optional): The method to reduce the loss.
+            Options are "none", "mean" and "sum". Defaults to "mean".
+        loss_weight (float, optional): The weight of loss.
+    """
+
+    def __init__(self, beta=1.0, reduction='mean', loss_weight=1.0):
+        super(SmoothL1Loss, self).__init__()
+        self.beta = beta
+        self.reduction = reduction
+        self.loss_weight = loss_weight
+
+    def forward(self,
+                pred,
+                target,
+                weight=None,
+                avg_factor=None,
+                reduction_override=None,
+                **kwargs):
+        """Forward function.
+
+        Args:
+            pred (torch.Tensor): The prediction.
+            target (torch.Tensor): The learning target of the prediction.
+            weight (torch.Tensor, optional): The weight of loss for each
+                prediction. Defaults to None.
+            avg_factor (int, optional): Average factor that is used to average
+                the loss. Defaults to None.
+            reduction_override (str, optional): The reduction method used to
+                override the original reduction method of the loss.
+                Defaults to None.
+        """
+        assert reduction_override in (None, 'none', 'mean', 'sum')
+        reduction = (
+            reduction_override if reduction_override else self.reduction)
+        loss_bbox = self.loss_weight * smooth_l1_loss(
+            pred,
+            target,
+            weight,
+            beta=self.beta,
+            reduction=reduction,
+            avg_factor=avg_factor,
+            **kwargs)
+        return loss_bbox
+
+
+@LOSSES.register_module()
+class L1Loss(nn.Module):
+    """L1 loss.
+
+    Args:
+        reduction (str, optional): The method to reduce the loss.
+            Options are "none", "mean" and "sum".
+        loss_weight (float, optional): The weight of loss.
+    """
+
+    def __init__(self, reduction='mean', loss_weight=1.0):
+        super(L1Loss, self).__init__()
+        self.reduction = reduction
+        self.loss_weight = loss_weight
+
+    def forward(self,
+                pred,
+                target,
+                weight=None,
+                avg_factor=None,
+                reduction_override=None):
+        """Forward function.
+
+        Args:
+            pred (torch.Tensor): The prediction.
+            target (torch.Tensor): The learning target of the prediction.
+            weight (torch.Tensor, optional): The weight of loss for each
+                prediction. Defaults to None.
+            avg_factor (int, optional): Average factor that is used to average
+                the loss. Defaults to None.
+            reduction_override (str, optional): The reduction method used to
+                override the original reduction method of the loss.
+                Defaults to None.
+        """
+        assert reduction_override in (None, 'none', 'mean', 'sum')
+        reduction = (
+            reduction_override if reduction_override else self.reduction)
+        loss_bbox = self.loss_weight * l1_loss(
+            pred, target, weight, reduction=reduction, avg_factor=avg_factor)
+        return loss_bbox

PyTorch/NLP/Conformer-main/mmdetection/mmdet/models/losses/utils.py

+import functools
+
+import torch.nn.functional as F
+
+
+def reduce_loss(loss, reduction):
+    """Reduce loss as specified.
+
+    Args:
+        loss (Tensor): Elementwise loss tensor.
+        reduction (str): Options are "none", "mean" and "sum".
+
+    Return:
+        Tensor: Reduced loss tensor.
+    """
+    reduction_enum = F._Reduction.get_enum(reduction)
+    # none: 0, elementwise_mean:1, sum: 2
+    if reduction_enum == 0:
+        return loss
+    elif reduction_enum == 1:
+        return loss.mean()
+    elif reduction_enum == 2:
+        return loss.sum()
+
+
+def weight_reduce_loss(loss, weight=None, reduction='mean', avg_factor=None):
+    """Apply element-wise weight and reduce loss.
+
+    Args:
+        loss (Tensor): Element-wise loss.
+        weight (Tensor): Element-wise weights.
+        reduction (str): Same as built-in losses of PyTorch.
+        avg_factor (float): Avarage factor when computing the mean of losses.
+
+    Returns:
+        Tensor: Processed loss values.
+    """
+    # if weight is specified, apply element-wise weight
+    if weight is not None:
+        loss = loss * weight
+
+    # if avg_factor is not specified, just reduce the loss
+    if avg_factor is None:
+        loss = reduce_loss(loss, reduction)
+    else:
+        # if reduction is mean, then average the loss by avg_factor
+        if reduction == 'mean':
+            loss = loss.sum() / avg_factor
+        # if reduction is 'none', then do nothing, otherwise raise an error
+        elif reduction != 'none':
+            raise ValueError('avg_factor can not be used with reduction="sum"')
+    return loss
+
+
+def weighted_loss(loss_func):
+    """Create a weighted version of a given loss function.
+
+    To use this decorator, the loss function must have the signature like
+    `loss_func(pred, target, **kwargs)`. The function only needs to compute
+    element-wise loss without any reduction. This decorator will add weight
+    and reduction arguments to the function. The decorated function will have
+    the signature like `loss_func(pred, target, weight=None, reduction='mean',
+    avg_factor=None, **kwargs)`.
+
+    :Example:
+
+    >>> import torch
+    >>> @weighted_loss
+    >>> def l1_loss(pred, target):
+    >>>     return (pred - target).abs()
+
+    >>> pred = torch.Tensor([0, 2, 3])
+    >>> target = torch.Tensor([1, 1, 1])
+    >>> weight = torch.Tensor([1, 0, 1])
+
+    >>> l1_loss(pred, target)
+    tensor(1.3333)
+    >>> l1_loss(pred, target, weight)
+    tensor(1.)
+    >>> l1_loss(pred, target, reduction='none')
+    tensor([1., 1., 2.])
+    >>> l1_loss(pred, target, weight, avg_factor=2)
+    tensor(1.5000)
+    """
+
+    @functools.wraps(loss_func)
+    def wrapper(pred,
+                target,
+                weight=None,
+                reduction='mean',
+                avg_factor=None,
+                **kwargs):
+        # get element-wise loss
+        loss = loss_func(pred, target, **kwargs)
+        loss = weight_reduce_loss(loss, weight, reduction, avg_factor)
+        return loss
+
+    return wrapper

PyTorch/NLP/Conformer-main/mmdetection/mmdet/models/losses/varifocal_loss.py

+import torch.nn as nn
+import torch.nn.functional as F
+
+from ..builder import LOSSES
+from .utils import weight_reduce_loss
+
+
+def varifocal_loss(pred,
+                   target,
+                   weight=None,
+                   alpha=0.75,
+                   gamma=2.0,
+                   iou_weighted=True,
+                   reduction='mean',
+                   avg_factor=None):
+    """`Varifocal Loss <https://arxiv.org/abs/2008.13367>`_
+
+    Args:
+        pred (torch.Tensor): The prediction with shape (N, C), C is the
+            number of classes
+        target (torch.Tensor): The learning target of the iou-aware
+            classification score with shape (N, C), C is the number of classes.
+        weight (torch.Tensor, optional): The weight of loss for each
+            prediction. Defaults to None.
+        alpha (float, optional): A balance factor for the negative part of
+            Varifocal Loss, which is different from the alpha of Focal Loss.
+            Defaults to 0.75.
+        gamma (float, optional): The gamma for calculating the modulating
+            factor. Defaults to 2.0.
+        iou_weighted (bool, optional): Whether to weight the loss of the
+            positive example with the iou target. Defaults to True.
+        reduction (str, optional): The method used to reduce the loss into
+            a scalar. Defaults to 'mean'. Options are "none", "mean" and
+            "sum".
+        avg_factor (int, optional): Average factor that is used to average
+            the loss. Defaults to None.
+    """
+    # pred and target should be of the same size
+    assert pred.size() == target.size()
+    pred_sigmoid = pred.sigmoid()
+    target = target.type_as(pred)
+    if iou_weighted:
+        focal_weight = target * (target > 0.0).float() + \
+            alpha * (pred_sigmoid - target).abs().pow(gamma) * \
+            (target <= 0.0).float()
+    else:
+        focal_weight = (target > 0.0).float() + \
+            alpha * (pred_sigmoid - target).abs().pow(gamma) * \
+            (target <= 0.0).float()
+    loss = F.binary_cross_entropy_with_logits(
+        pred, target, reduction='none') * focal_weight
+    loss = weight_reduce_loss(loss, weight, reduction, avg_factor)
+    return loss
+
+
+@LOSSES.register_module()
+class VarifocalLoss(nn.Module):
+
+    def __init__(self,
+                 use_sigmoid=True,
+                 alpha=0.75,
+                 gamma=2.0,
+                 iou_weighted=True,
+                 reduction='mean',
+                 loss_weight=1.0):
+        """`Varifocal Loss <https://arxiv.org/abs/2008.13367>`_
+
+        Args:
+            use_sigmoid (bool, optional): Whether the prediction is
+                used for sigmoid or softmax. Defaults to True.
+            alpha (float, optional): A balance factor for the negative part of
+                Varifocal Loss, which is different from the alpha of Focal
+                Loss. Defaults to 0.75.
+            gamma (float, optional): The gamma for calculating the modulating
+                factor. Defaults to 2.0.
+            iou_weighted (bool, optional): Whether to weight the loss of the
+                positive examples with the iou target. Defaults to True.
+            reduction (str, optional): The method used to reduce the loss into
+                a scalar. Defaults to 'mean'. Options are "none", "mean" and
+                "sum".
+            loss_weight (float, optional): Weight of loss. Defaults to 1.0.
+        """
+        super(VarifocalLoss, self).__init__()
+        assert use_sigmoid is True, \
+            'Only sigmoid varifocal loss supported now.'
+        assert alpha >= 0.0
+        self.use_sigmoid = use_sigmoid
+        self.alpha = alpha
+        self.gamma = gamma
+        self.iou_weighted = iou_weighted
+        self.reduction = reduction
+        self.loss_weight = loss_weight
+
+    def forward(self,
+                pred,
+                target,
+                weight=None,
+                avg_factor=None,
+                reduction_override=None):
+        """Forward function.
+
+        Args:
+            pred (torch.Tensor): The prediction.
+            target (torch.Tensor): The learning target of the prediction.
+            weight (torch.Tensor, optional): The weight of loss for each
+                prediction. Defaults to None.
+            avg_factor (int, optional): Average factor that is used to average
+                the loss. Defaults to None.
+            reduction_override (str, optional): The reduction method used to
+                override the original reduction method of the loss.
+                Options are "none", "mean" and "sum".
+
+        Returns:
+            torch.Tensor: The calculated loss
+        """
+        assert reduction_override in (None, 'none', 'mean', 'sum')
+        reduction = (
+            reduction_override if reduction_override else self.reduction)
+        if self.use_sigmoid:
+            loss_cls = self.loss_weight * varifocal_loss(
+                pred,
+                target,
+                weight,
+                alpha=self.alpha,
+                gamma=self.gamma,
+                iou_weighted=self.iou_weighted,
+                reduction=reduction,
+                avg_factor=avg_factor)
+        else:
+            raise NotImplementedError
+        return loss_cls

PyTorch/NLP/Conformer-main/mmdetection/mmdet/models/necks/__init__.py

+from .bfp import BFP
+from .channel_mapper import ChannelMapper
+from .fpn import FPN
+from .fpn_carafe import FPN_CARAFE
+from .hrfpn import HRFPN
+from .nas_fpn import NASFPN
+from .nasfcos_fpn import NASFCOS_FPN
+from .pafpn import PAFPN
+from .rfp import RFP
+from .yolo_neck import YOLOV3Neck
+
+__all__ = [
+    'FPN', 'BFP', 'ChannelMapper', 'HRFPN', 'NASFPN', 'FPN_CARAFE', 'PAFPN',
+    'NASFCOS_FPN', 'RFP', 'YOLOV3Neck'
+]

PyTorch/NLP/Conformer-main/mmdetection/mmdet/models/necks/bfp.py

+import torch.nn as nn
+import torch.nn.functional as F
+from mmcv.cnn import ConvModule, xavier_init
+from mmcv.cnn.bricks import NonLocal2d
+
+from ..builder import NECKS
+
+
+@NECKS.register_module()
+class BFP(nn.Module):
+    """BFP (Balanced Feature Pyrmamids)
+
+    BFP takes multi-level features as inputs and gather them into a single one,
+    then refine the gathered feature and scatter the refined results to
+    multi-level features. This module is used in Libra R-CNN (CVPR 2019), see
+    the paper `Libra R-CNN: Towards Balanced Learning for Object Detection
+    <https://arxiv.org/abs/1904.02701>`_ for details.
+
+    Args:
+        in_channels (int): Number of input channels (feature maps of all levels
+            should have the same channels).
+        num_levels (int): Number of input feature levels.
+        conv_cfg (dict): The config dict for convolution layers.
+        norm_cfg (dict): The config dict for normalization layers.
+        refine_level (int): Index of integration and refine level of BSF in
+            multi-level features from bottom to top.
+        refine_type (str): Type of the refine op, currently support
+            [None, 'conv', 'non_local'].
+    """
+
+    def __init__(self,
+                 in_channels,
+                 num_levels,
+                 refine_level=2,
+                 refine_type=None,
+                 conv_cfg=None,
+                 norm_cfg=None):
+        super(BFP, self).__init__()
+        assert refine_type in [None, 'conv', 'non_local']
+
+        self.in_channels = in_channels
+        self.num_levels = num_levels
+        self.conv_cfg = conv_cfg
+        self.norm_cfg = norm_cfg
+
+        self.refine_level = refine_level
+        self.refine_type = refine_type
+        assert 0 <= self.refine_level < self.num_levels
+
+        if self.refine_type == 'conv':
+            self.refine = ConvModule(
+                self.in_channels,
+                self.in_channels,
+                3,
+                padding=1,
+                conv_cfg=self.conv_cfg,
+                norm_cfg=self.norm_cfg)
+        elif self.refine_type == 'non_local':
+            self.refine = NonLocal2d(
+                self.in_channels,
+                reduction=1,
+                use_scale=False,
+                conv_cfg=self.conv_cfg,
+                norm_cfg=self.norm_cfg)
+
+    def init_weights(self):
+        """Initialize the weights of FPN module."""
+        for m in self.modules():
+            if isinstance(m, nn.Conv2d):
+                xavier_init(m, distribution='uniform')
+
+    def forward(self, inputs):
+        """Forward function."""
+        assert len(inputs) == self.num_levels
+
+        # step 1: gather multi-level features by resize and average
+        feats = []
+        gather_size = inputs[self.refine_level].size()[2:]
+        for i in range(self.num_levels):
+            if i < self.refine_level:
+                gathered = F.adaptive_max_pool2d(
+                    inputs[i], output_size=gather_size)
+            else:
+                gathered = F.interpolate(
+                    inputs[i], size=gather_size, mode='nearest')
+            feats.append(gathered)
+
+        bsf = sum(feats) / len(feats)
+
+        # step 2: refine gathered features
+        if self.refine_type is not None:
+            bsf = self.refine(bsf)
+
+        # step 3: scatter refined features to multi-levels by a residual path
+        outs = []
+        for i in range(self.num_levels):
+            out_size = inputs[i].size()[2:]
+            if i < self.refine_level:
+                residual = F.interpolate(bsf, size=out_size, mode='nearest')
+            else:
+                residual = F.adaptive_max_pool2d(bsf, output_size=out_size)
+            outs.append(residual + inputs[i])
+
+        return tuple(outs)

PyTorch/NLP/Conformer-main/mmdetection/mmdet/models/necks/channel_mapper.py

+import torch.nn as nn
+from mmcv.cnn import ConvModule, xavier_init
+
+from ..builder import NECKS
+
+
+@NECKS.register_module()
+class ChannelMapper(nn.Module):
+    r"""Channel Mapper to reduce/increase channels of backbone features.
+
+    This is used to reduce/increase channels of backbone features.
+
+    Args:
+        in_channels (List[int]): Number of input channels per scale.
+        out_channels (int): Number of output channels (used at each scale).
+        kernel_size (int, optional): kernel_size for reducing channels (used
+            at each scale). Default: 3.
+        conv_cfg (dict, optional): Config dict for convolution layer.
+            Default: None.
+        norm_cfg (dict, optional): Config dict for normalization layer.
+            Default: None.
+        act_cfg (dict, optional): Config dict for activation layer in
+            ConvModule. Default: dict(type='ReLU').
+
+    Example:
+        >>> import torch
+        >>> in_channels = [2, 3, 5, 7]
+        >>> scales = [340, 170, 84, 43]
+        >>> inputs = [torch.rand(1, c, s, s)
+        ...           for c, s in zip(in_channels, scales)]
+        >>> self = ChannelMapper(in_channels, 11, 3).eval()
+        >>> outputs = self.forward(inputs)
+        >>> for i in range(len(outputs)):
+        ...     print(f'outputs[{i}].shape = {outputs[i].shape}')
+        outputs[0].shape = torch.Size([1, 11, 340, 340])
+        outputs[1].shape = torch.Size([1, 11, 170, 170])
+        outputs[2].shape = torch.Size([1, 11, 84, 84])
+        outputs[3].shape = torch.Size([1, 11, 43, 43])
+    """
+
+    def __init__(self,
+                 in_channels,
+                 out_channels,
+                 kernel_size=3,
+                 conv_cfg=None,
+                 norm_cfg=None,
+                 act_cfg=dict(type='ReLU')):
+        super(ChannelMapper, self).__init__()
+        assert isinstance(in_channels, list)
+
+        self.convs = nn.ModuleList()
+        for in_channel in in_channels:
+            self.convs.append(
+                ConvModule(
+                    in_channel,
+                    out_channels,
+                    kernel_size,
+                    padding=(kernel_size - 1) // 2,
+                    conv_cfg=conv_cfg,
+                    norm_cfg=norm_cfg,
+                    act_cfg=act_cfg))
+
+    # default init_weights for conv(msra) and norm in ConvModule
+    def init_weights(self):
+        """Initialize the weights of ChannelMapper module."""
+        for m in self.modules():
+            if isinstance(m, nn.Conv2d):
+                xavier_init(m, distribution='uniform')
+
+    def forward(self, inputs):
+        """Forward function."""
+        assert len(inputs) == len(self.convs)
+        outs = [self.convs[i](inputs[i]) for i in range(len(inputs))]
+        return tuple(outs)

PyTorch/NLP/Conformer-main/mmdetection/mmdet/models/necks/fpn.py

+import warnings
+
+import torch.nn as nn
+import torch.nn.functional as F
+from mmcv.cnn import ConvModule, xavier_init
+from mmcv.runner import auto_fp16
+
+from ..builder import NECKS
+
+
+@NECKS.register_module()
+class FPN(nn.Module):
+    r"""Feature Pyramid Network.
+
+    This is an implementation of paper `Feature Pyramid Networks for Object
+    Detection <https://arxiv.org/abs/1612.03144>`_.
+
+    Args:
+        in_channels (List[int]): Number of input channels per scale.
+        out_channels (int): Number of output channels (used at each scale)
+        num_outs (int): Number of output scales.
+        start_level (int): Index of the start input backbone level used to
+            build the feature pyramid. Default: 0.
+        end_level (int): Index of the end input backbone level (exclusive) to
+            build the feature pyramid. Default: -1, which means the last level.
+        add_extra_convs (bool | str): If bool, it decides whether to add conv
+            layers on top of the original feature maps. Default to False.
+            If True, its actual mode is specified by `extra_convs_on_inputs`.
+            If str, it specifies the source feature map of the extra convs.
+            Only the following options are allowed
+
+            - 'on_input': Last feat map of neck inputs (i.e. backbone feature).
+            - 'on_lateral':  Last feature map after lateral convs.
+            - 'on_output': The last output feature map after fpn convs.
+        extra_convs_on_inputs (bool, deprecated): Whether to apply extra convs
+            on the original feature from the backbone. If True,
+            it is equivalent to `add_extra_convs='on_input'`. If False, it is
+            equivalent to set `add_extra_convs='on_output'`. Default to True.
+        relu_before_extra_convs (bool): Whether to apply relu before the extra
+            conv. Default: False.
+        no_norm_on_lateral (bool): Whether to apply norm on lateral.
+            Default: False.
+        conv_cfg (dict): Config dict for convolution layer. Default: None.
+        norm_cfg (dict): Config dict for normalization layer. Default: None.
+        act_cfg (str): Config dict for activation layer in ConvModule.
+            Default: None.
+        upsample_cfg (dict): Config dict for interpolate layer.
+            Default: `dict(mode='nearest')`
+
+    Example:
+        >>> import torch
+        >>> in_channels = [2, 3, 5, 7]
+        >>> scales = [340, 170, 84, 43]
+        >>> inputs = [torch.rand(1, c, s, s)
+        ...           for c, s in zip(in_channels, scales)]
+        >>> self = FPN(in_channels, 11, len(in_channels)).eval()
+        >>> outputs = self.forward(inputs)
+        >>> for i in range(len(outputs)):
+        ...     print(f'outputs[{i}].shape = {outputs[i].shape}')
+        outputs[0].shape = torch.Size([1, 11, 340, 340])
+        outputs[1].shape = torch.Size([1, 11, 170, 170])
+        outputs[2].shape = torch.Size([1, 11, 84, 84])
+        outputs[3].shape = torch.Size([1, 11, 43, 43])
+    """
+
+    def __init__(self,
+                 in_channels,
+                 out_channels,
+                 num_outs,
+                 start_level=0,
+                 end_level=-1,
+                 add_extra_convs=False,
+                 extra_convs_on_inputs=True,
+                 relu_before_extra_convs=False,
+                 no_norm_on_lateral=False,
+                 conv_cfg=None,
+                 norm_cfg=None,
+                 act_cfg=None,
+                 upsample_cfg=dict(mode='nearest')):
+        super(FPN, self).__init__()
+        assert isinstance(in_channels, list)
+        self.in_channels = in_channels
+        self.out_channels = out_channels
+        self.num_ins = len(in_channels)
+        self.num_outs = num_outs
+        self.relu_before_extra_convs = relu_before_extra_convs
+        self.no_norm_on_lateral = no_norm_on_lateral
+        self.fp16_enabled = False
+        self.upsample_cfg = upsample_cfg.copy()
+
+        if end_level == -1:
+            self.backbone_end_level = self.num_ins
+            assert num_outs >= self.num_ins - start_level
+        else:
+            # if end_level < inputs, no extra level is allowed
+            self.backbone_end_level = end_level
+            assert end_level <= len(in_channels)
+            assert num_outs == end_level - start_level
+        self.start_level = start_level
+        self.end_level = end_level
+        self.add_extra_convs = add_extra_convs
+        assert isinstance(add_extra_convs, (str, bool))
+        if isinstance(add_extra_convs, str):
+            # Extra_convs_source choices: 'on_input', 'on_lateral', 'on_output'
+            assert add_extra_convs in ('on_input', 'on_lateral', 'on_output')
+        elif add_extra_convs:  # True
+            if extra_convs_on_inputs:
+                # TODO: deprecate `extra_convs_on_inputs`
+                warnings.simplefilter('once')
+                warnings.warn(
+                    '"extra_convs_on_inputs" will be deprecated in v2.9.0,'
+                    'Please use "add_extra_convs"', DeprecationWarning)
+                self.add_extra_convs = 'on_input'
+            else:
+                self.add_extra_convs = 'on_output'
+
+        self.lateral_convs = nn.ModuleList()
+        self.fpn_convs = nn.ModuleList()
+
+        for i in range(self.start_level, self.backbone_end_level):
+            l_conv = ConvModule(
+                in_channels[i],
+                out_channels,
+                1,
+                conv_cfg=conv_cfg,
+                norm_cfg=norm_cfg if not self.no_norm_on_lateral else None,
+                act_cfg=act_cfg,
+                inplace=False)
+            fpn_conv = ConvModule(
+                out_channels,
+                out_channels,
+                3,
+                padding=1,
+                conv_cfg=conv_cfg,
+                norm_cfg=norm_cfg,
+                act_cfg=act_cfg,
+                inplace=False)
+
+            self.lateral_convs.append(l_conv)
+            self.fpn_convs.append(fpn_conv)
+
+        # add extra conv layers (e.g., RetinaNet)
+        extra_levels = num_outs - self.backbone_end_level + self.start_level
+        if self.add_extra_convs and extra_levels >= 1:
+            for i in range(extra_levels):
+                if i == 0 and self.add_extra_convs == 'on_input':
+                    in_channels = self.in_channels[self.backbone_end_level - 1]
+                else:
+                    in_channels = out_channels
+                extra_fpn_conv = ConvModule(
+                    in_channels,
+                    out_channels,
+                    3,
+                    stride=2,
+                    padding=1,
+                    conv_cfg=conv_cfg,
+                    norm_cfg=norm_cfg,
+                    act_cfg=act_cfg,
+                    inplace=False)
+                self.fpn_convs.append(extra_fpn_conv)
+
+    # default init_weights for conv(msra) and norm in ConvModule
+    def init_weights(self):
+        """Initialize the weights of FPN module."""
+        for m in self.modules():
+            if isinstance(m, nn.Conv2d):
+                xavier_init(m, distribution='uniform')
+
+    @auto_fp16()
+    def forward(self, inputs):
+        """Forward function."""
+        assert len(inputs) == len(self.in_channels)
+
+        # build laterals
+        laterals = [
+            lateral_conv(inputs[i + self.start_level])
+            for i, lateral_conv in enumerate(self.lateral_convs)
+        ]
+
+        # build top-down path
+        used_backbone_levels = len(laterals)
+        for i in range(used_backbone_levels - 1, 0, -1):
+            # In some cases, fixing `scale factor` (e.g. 2) is preferred, but
+            #  it cannot co-exist with `size` in `F.interpolate`.
+            if 'scale_factor' in self.upsample_cfg:
+                laterals[i - 1] += F.interpolate(laterals[i],
+                                                 **self.upsample_cfg)
+            else:
+                prev_shape = laterals[i - 1].shape[2:]
+                laterals[i - 1] += F.interpolate(
+                    laterals[i], size=prev_shape, **self.upsample_cfg)
+
+        # build outputs
+        # part 1: from original levels
+        outs = [
+            self.fpn_convs[i](laterals[i]) for i in range(used_backbone_levels)
+        ]
+        # part 2: add extra levels
+        if self.num_outs > len(outs):
+            # use max pool to get more levels on top of outputs
+            # (e.g., Faster R-CNN, Mask R-CNN)
+            if not self.add_extra_convs:
+                for i in range(self.num_outs - used_backbone_levels):
+                    outs.append(F.max_pool2d(outs[-1], 1, stride=2))
+            # add conv layers on top of original feature maps (RetinaNet)
+            else:
+                if self.add_extra_convs == 'on_input':
+                    extra_source = inputs[self.backbone_end_level - 1]
+                elif self.add_extra_convs == 'on_lateral':
+                    extra_source = laterals[-1]
+                elif self.add_extra_convs == 'on_output':
+                    extra_source = outs[-1]
+                else:
+                    raise NotImplementedError
+                outs.append(self.fpn_convs[used_backbone_levels](extra_source))
+                for i in range(used_backbone_levels + 1, self.num_outs):
+                    if self.relu_before_extra_convs:
+                        outs.append(self.fpn_convs[i](F.relu(outs[-1])))
+                    else:
+                        outs.append(self.fpn_convs[i](outs[-1]))
+        return tuple(outs)

PyTorch/NLP/Conformer-main/mmdetection/mmdet/models/necks/fpn_carafe.py

+import torch.nn as nn
+from mmcv.cnn import ConvModule, build_upsample_layer, xavier_init
+from mmcv.ops.carafe import CARAFEPack
+
+from ..builder import NECKS
+
+
+@NECKS.register_module()
+class FPN_CARAFE(nn.Module):
+    """FPN_CARAFE is a more flexible implementation of FPN. It allows more
+    choice for upsample methods during the top-down pathway.
+
+    It can reproduce the preformance of ICCV 2019 paper
+    CARAFE: Content-Aware ReAssembly of FEatures
+    Please refer to https://arxiv.org/abs/1905.02188 for more details.
+
+    Args:
+        in_channels (list[int]): Number of channels for each input feature map.
+        out_channels (int): Output channels of feature pyramids.
+        num_outs (int): Number of output stages.
+        start_level (int): Start level of feature pyramids.
+            (Default: 0)
+        end_level (int): End level of feature pyramids.
+            (Default: -1 indicates the last level).
+        norm_cfg (dict): Dictionary to construct and config norm layer.
+        activate (str): Type of activation function in ConvModule
+            (Default: None indicates w/o activation).
+        order (dict): Order of components in ConvModule.
+        upsample (str): Type of upsample layer.
+        upsample_cfg (dict): Dictionary to construct and config upsample layer.
+    """
+
+    def __init__(self,
+                 in_channels,
+                 out_channels,
+                 num_outs,
+                 start_level=0,
+                 end_level=-1,
+                 norm_cfg=None,
+                 act_cfg=None,
+                 order=('conv', 'norm', 'act'),
+                 upsample_cfg=dict(
+                     type='carafe',
+                     up_kernel=5,
+                     up_group=1,
+                     encoder_kernel=3,
+                     encoder_dilation=1)):
+        super(FPN_CARAFE, self).__init__()
+        assert isinstance(in_channels, list)
+        self.in_channels = in_channels
+        self.out_channels = out_channels
+        self.num_ins = len(in_channels)
+        self.num_outs = num_outs
+        self.norm_cfg = norm_cfg
+        self.act_cfg = act_cfg
+        self.with_bias = norm_cfg is None
+        self.upsample_cfg = upsample_cfg.copy()
+        self.upsample = self.upsample_cfg.get('type')
+        self.relu = nn.ReLU(inplace=False)
+
+        self.order = order
+        assert order in [('conv', 'norm', 'act'), ('act', 'conv', 'norm')]
+
+        assert self.upsample in [
+            'nearest', 'bilinear', 'deconv', 'pixel_shuffle', 'carafe', None
+        ]
+        if self.upsample in ['deconv', 'pixel_shuffle']:
+            assert hasattr(
+                self.upsample_cfg,
+                'upsample_kernel') and self.upsample_cfg.upsample_kernel > 0
+            self.upsample_kernel = self.upsample_cfg.pop('upsample_kernel')
+
+        if end_level == -1:
+            self.backbone_end_level = self.num_ins
+            assert num_outs >= self.num_ins - start_level
+        else:
+            # if end_level < inputs, no extra level is allowed
+            self.backbone_end_level = end_level
+            assert end_level <= len(in_channels)
+            assert num_outs == end_level - start_level
+        self.start_level = start_level
+        self.end_level = end_level
+
+        self.lateral_convs = nn.ModuleList()
+        self.fpn_convs = nn.ModuleList()
+        self.upsample_modules = nn.ModuleList()
+
+        for i in range(self.start_level, self.backbone_end_level):
+            l_conv = ConvModule(
+                in_channels[i],
+                out_channels,
+                1,
+                norm_cfg=norm_cfg,
+                bias=self.with_bias,
+                act_cfg=act_cfg,
+                inplace=False,
+                order=self.order)
+            fpn_conv = ConvModule(
+                out_channels,
+                out_channels,
+                3,
+                padding=1,
+                norm_cfg=self.norm_cfg,
+                bias=self.with_bias,
+                act_cfg=act_cfg,
+                inplace=False,
+                order=self.order)
+            if i != self.backbone_end_level - 1:
+                upsample_cfg_ = self.upsample_cfg.copy()
+                if self.upsample == 'deconv':
+                    upsample_cfg_.update(
+                        in_channels=out_channels,
+                        out_channels=out_channels,
+                        kernel_size=self.upsample_kernel,
+                        stride=2,
+                        padding=(self.upsample_kernel - 1) // 2,
+                        output_padding=(self.upsample_kernel - 1) // 2)
+                elif self.upsample == 'pixel_shuffle':
+                    upsample_cfg_.update(
+                        in_channels=out_channels,
+                        out_channels=out_channels,
+                        scale_factor=2,
+                        upsample_kernel=self.upsample_kernel)
+                elif self.upsample == 'carafe':
+                    upsample_cfg_.update(channels=out_channels, scale_factor=2)
+                else:
+                    # suppress warnings
+                    align_corners = (None
+                                     if self.upsample == 'nearest' else False)
+                    upsample_cfg_.update(
+                        scale_factor=2,
+                        mode=self.upsample,
+                        align_corners=align_corners)
+                upsample_module = build_upsample_layer(upsample_cfg_)
+                self.upsample_modules.append(upsample_module)
+            self.lateral_convs.append(l_conv)
+            self.fpn_convs.append(fpn_conv)
+
+        # add extra conv layers (e.g., RetinaNet)
+        extra_out_levels = (
+            num_outs - self.backbone_end_level + self.start_level)
+        if extra_out_levels >= 1:
+            for i in range(extra_out_levels):
+                in_channels = (
+                    self.in_channels[self.backbone_end_level -
+                                     1] if i == 0 else out_channels)
+                extra_l_conv = ConvModule(
+                    in_channels,
+                    out_channels,
+                    3,
+                    stride=2,
+                    padding=1,
+                    norm_cfg=norm_cfg,
+                    bias=self.with_bias,
+                    act_cfg=act_cfg,
+                    inplace=False,
+                    order=self.order)
+                if self.upsample == 'deconv':
+                    upsampler_cfg_ = dict(
+                        in_channels=out_channels,
+                        out_channels=out_channels,
+                        kernel_size=self.upsample_kernel,
+                        stride=2,
+                        padding=(self.upsample_kernel - 1) // 2,
+                        output_padding=(self.upsample_kernel - 1) // 2)
+                elif self.upsample == 'pixel_shuffle':
+                    upsampler_cfg_ = dict(
+                        in_channels=out_channels,
+                        out_channels=out_channels,
+                        scale_factor=2,
+                        upsample_kernel=self.upsample_kernel)
+                elif self.upsample == 'carafe':
+                    upsampler_cfg_ = dict(
+                        channels=out_channels,
+                        scale_factor=2,
+                        **self.upsample_cfg)
+                else:
+                    # suppress warnings
+                    align_corners = (None
+                                     if self.upsample == 'nearest' else False)
+                    upsampler_cfg_ = dict(
+                        scale_factor=2,
+                        mode=self.upsample,
+                        align_corners=align_corners)
+                upsampler_cfg_['type'] = self.upsample
+                upsample_module = build_upsample_layer(upsampler_cfg_)
+                extra_fpn_conv = ConvModule(
+                    out_channels,
+                    out_channels,
+                    3,
+                    padding=1,
+                    norm_cfg=self.norm_cfg,
+                    bias=self.with_bias,
+                    act_cfg=act_cfg,
+                    inplace=False,
+                    order=self.order)
+                self.upsample_modules.append(upsample_module)
+                self.fpn_convs.append(extra_fpn_conv)
+                self.lateral_convs.append(extra_l_conv)
+
+    # default init_weights for conv(msra) and norm in ConvModule
+    def init_weights(self):
+        """Initialize the weights of module."""
+        for m in self.modules():
+            if isinstance(m, (nn.Conv2d, nn.ConvTranspose2d)):
+                xavier_init(m, distribution='uniform')
+        for m in self.modules():
+            if isinstance(m, CARAFEPack):
+                m.init_weights()
+
+    def slice_as(self, src, dst):
+        """Slice ``src`` as ``dst``
+
+        Note:
+            ``src`` should have the same or larger size than ``dst``.
+
+        Args:
+            src (torch.Tensor): Tensors to be sliced.
+            dst (torch.Tensor): ``src`` will be sliced to have the same
+                size as ``dst``.
+
+        Returns:
+            torch.Tensor: Sliced tensor.
+        """
+        assert (src.size(2) >= dst.size(2)) and (src.size(3) >= dst.size(3))
+        if src.size(2) == dst.size(2) and src.size(3) == dst.size(3):
+            return src
+        else:
+            return src[:, :, :dst.size(2), :dst.size(3)]
+
+    def tensor_add(self, a, b):
+        """Add tensors ``a`` and ``b`` that might have different sizes."""
+        if a.size() == b.size():
+            c = a + b
+        else:
+            c = a + self.slice_as(b, a)
+        return c
+
+    def forward(self, inputs):
+        """Forward function."""
+        assert len(inputs) == len(self.in_channels)
+
+        # build laterals
+        laterals = []
+        for i, lateral_conv in enumerate(self.lateral_convs):
+            if i <= self.backbone_end_level - self.start_level:
+                input = inputs[min(i + self.start_level, len(inputs) - 1)]
+            else:
+                input = laterals[-1]
+            lateral = lateral_conv(input)
+            laterals.append(lateral)
+
+        # build top-down path
+        for i in range(len(laterals) - 1, 0, -1):
+            if self.upsample is not None:
+                upsample_feat = self.upsample_modules[i - 1](laterals[i])
+            else:
+                upsample_feat = laterals[i]
+            laterals[i - 1] = self.tensor_add(laterals[i - 1], upsample_feat)
+
+        # build outputs
+        num_conv_outs = len(self.fpn_convs)
+        outs = []
+        for i in range(num_conv_outs):
+            out = self.fpn_convs[i](laterals[i])
+            outs.append(out)
+        return tuple(outs)

PyTorch/NLP/Conformer-main/mmdetection/mmdet/models/necks/hrfpn.py

+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+from mmcv.cnn import ConvModule, caffe2_xavier_init
+from torch.utils.checkpoint import checkpoint
+
+from ..builder import NECKS
+
+
+@NECKS.register_module()
+class HRFPN(nn.Module):
+    """HRFPN (High Resolution Feature Pyrmamids)
+
+    paper: `High-Resolution Representations for Labeling Pixels and Regions
+    <https://arxiv.org/abs/1904.04514>`_.
+
+    Args:
+        in_channels (list): number of channels for each branch.
+        out_channels (int): output channels of feature pyramids.
+        num_outs (int): number of output stages.
+        pooling_type (str): pooling for generating feature pyramids
+            from {MAX, AVG}.
+        conv_cfg (dict): dictionary to construct and config conv layer.
+        norm_cfg (dict): dictionary to construct and config norm layer.
+        with_cp  (bool): Use checkpoint or not. Using checkpoint will save some
+            memory while slowing down the training speed.
+        stride (int): stride of 3x3 convolutional layers
+    """
+
+    def __init__(self,
+                 in_channels,
+                 out_channels,
+                 num_outs=5,
+                 pooling_type='AVG',
+                 conv_cfg=None,
+                 norm_cfg=None,
+                 with_cp=False,
+                 stride=1):
+        super(HRFPN, self).__init__()
+        assert isinstance(in_channels, list)
+        self.in_channels = in_channels
+        self.out_channels = out_channels
+        self.num_ins = len(in_channels)
+        self.num_outs = num_outs
+        self.with_cp = with_cp
+        self.conv_cfg = conv_cfg
+        self.norm_cfg = norm_cfg
+
+        self.reduction_conv = ConvModule(
+            sum(in_channels),
+            out_channels,
+            kernel_size=1,
+            conv_cfg=self.conv_cfg,
+            act_cfg=None)
+
+        self.fpn_convs = nn.ModuleList()
+        for i in range(self.num_outs):
+            self.fpn_convs.append(
+                ConvModule(
+                    out_channels,
+                    out_channels,
+                    kernel_size=3,
+                    padding=1,
+                    stride=stride,
+                    conv_cfg=self.conv_cfg,
+                    act_cfg=None))
+
+        if pooling_type == 'MAX':
+            self.pooling = F.max_pool2d
+        else:
+            self.pooling = F.avg_pool2d
+
+    def init_weights(self):
+        """Initialize the weights of module."""
+        for m in self.modules():
+            if isinstance(m, nn.Conv2d):
+                caffe2_xavier_init(m)
+
+    def forward(self, inputs):
+        """Forward function."""
+        assert len(inputs) == self.num_ins
+        outs = [inputs[0]]
+        for i in range(1, self.num_ins):
+            outs.append(
+                F.interpolate(inputs[i], scale_factor=2**i, mode='bilinear'))
+        out = torch.cat(outs, dim=1)
+        if out.requires_grad and self.with_cp:
+            out = checkpoint(self.reduction_conv, out)
+        else:
+            out = self.reduction_conv(out)
+        outs = [out]
+        for i in range(1, self.num_outs):
+            outs.append(self.pooling(out, kernel_size=2**i, stride=2**i))
+        outputs = []
+
+        for i in range(self.num_outs):
+            if outs[i].requires_grad and self.with_cp:
+                tmp_out = checkpoint(self.fpn_convs[i], outs[i])
+            else:
+                tmp_out = self.fpn_convs[i](outs[i])
+            outputs.append(tmp_out)
+        return tuple(outputs)

PyTorch/NLP/Conformer-main/mmdetection/mmdet/models/necks/nas_fpn.py

+import torch.nn as nn
+from mmcv.cnn import ConvModule, caffe2_xavier_init
+from mmcv.ops.merge_cells import GlobalPoolingCell, SumCell
+
+from ..builder import NECKS
+
+
+@NECKS.register_module()
+class NASFPN(nn.Module):
+    """NAS-FPN.
+
+    Implementation of `NAS-FPN: Learning Scalable Feature Pyramid Architecture
+    for Object Detection <https://arxiv.org/abs/1904.07392>`_
+
+    Args:
+        in_channels (List[int]): Number of input channels per scale.
+        out_channels (int): Number of output channels (used at each scale)
+        num_outs (int): Number of output scales.
+        stack_times (int): The number of times the pyramid architecture will
+            be stacked.
+        start_level (int): Index of the start input backbone level used to
+            build the feature pyramid. Default: 0.
+        end_level (int): Index of the end input backbone level (exclusive) to
+            build the feature pyramid. Default: -1, which means the last level.
+        add_extra_convs (bool): It decides whether to add conv
+            layers on top of the original feature maps. Default to False.
+            If True, its actual mode is specified by `extra_convs_on_inputs`.
+    """
+
+    def __init__(self,
+                 in_channels,
+                 out_channels,
+                 num_outs,
+                 stack_times,
+                 start_level=0,
+                 end_level=-1,
+                 add_extra_convs=False,
+                 norm_cfg=None):
+        super(NASFPN, self).__init__()
+        assert isinstance(in_channels, list)
+        self.in_channels = in_channels
+        self.out_channels = out_channels
+        self.num_ins = len(in_channels)  # num of input feature levels
+        self.num_outs = num_outs  # num of output feature levels
+        self.stack_times = stack_times
+        self.norm_cfg = norm_cfg
+
+        if end_level == -1:
+            self.backbone_end_level = self.num_ins
+            assert num_outs >= self.num_ins - start_level
+        else:
+            # if end_level < inputs, no extra level is allowed
+            self.backbone_end_level = end_level
+            assert end_level <= len(in_channels)
+            assert num_outs == end_level - start_level
+        self.start_level = start_level
+        self.end_level = end_level
+        self.add_extra_convs = add_extra_convs
+
+        # add lateral connections
+        self.lateral_convs = nn.ModuleList()
+        for i in range(self.start_level, self.backbone_end_level):
+            l_conv = ConvModule(
+                in_channels[i],
+                out_channels,
+                1,
+                norm_cfg=norm_cfg,
+                act_cfg=None)
+            self.lateral_convs.append(l_conv)
+
+        # add extra downsample layers (stride-2 pooling or conv)
+        extra_levels = num_outs - self.backbone_end_level + self.start_level
+        self.extra_downsamples = nn.ModuleList()
+        for i in range(extra_levels):
+            extra_conv = ConvModule(
+                out_channels, out_channels, 1, norm_cfg=norm_cfg, act_cfg=None)
+            self.extra_downsamples.append(
+                nn.Sequential(extra_conv, nn.MaxPool2d(2, 2)))
+
+        # add NAS FPN connections
+        self.fpn_stages = nn.ModuleList()
+        for _ in range(self.stack_times):
+            stage = nn.ModuleDict()
+            # gp(p6, p4) -> p4_1
+            stage['gp_64_4'] = GlobalPoolingCell(
+                in_channels=out_channels,
+                out_channels=out_channels,
+                out_norm_cfg=norm_cfg)
+            # sum(p4_1, p4) -> p4_2
+            stage['sum_44_4'] = SumCell(
+                in_channels=out_channels,
+                out_channels=out_channels,
+                out_norm_cfg=norm_cfg)
+            # sum(p4_2, p3) -> p3_out
+            stage['sum_43_3'] = SumCell(
+                in_channels=out_channels,
+                out_channels=out_channels,
+                out_norm_cfg=norm_cfg)
+            # sum(p3_out, p4_2) -> p4_out
+            stage['sum_34_4'] = SumCell(
+                in_channels=out_channels,
+                out_channels=out_channels,
+                out_norm_cfg=norm_cfg)
+            # sum(p5, gp(p4_out, p3_out)) -> p5_out
+            stage['gp_43_5'] = GlobalPoolingCell(with_out_conv=False)
+            stage['sum_55_5'] = SumCell(
+                in_channels=out_channels,
+                out_channels=out_channels,
+                out_norm_cfg=norm_cfg)
+            # sum(p7, gp(p5_out, p4_2)) -> p7_out
+            stage['gp_54_7'] = GlobalPoolingCell(with_out_conv=False)
+            stage['sum_77_7'] = SumCell(
+                in_channels=out_channels,
+                out_channels=out_channels,
+                out_norm_cfg=norm_cfg)
+            # gp(p7_out, p5_out) -> p6_out
+            stage['gp_75_6'] = GlobalPoolingCell(
+                in_channels=out_channels,
+                out_channels=out_channels,
+                out_norm_cfg=norm_cfg)
+            self.fpn_stages.append(stage)
+
+    def init_weights(self):
+        """Initialize the weights of module."""
+        for m in self.modules():
+            if isinstance(m, nn.Conv2d):
+                caffe2_xavier_init(m)
+
+    def forward(self, inputs):
+        """Forward function."""
+        # build P3-P5
+        feats = [
+            lateral_conv(inputs[i + self.start_level])
+            for i, lateral_conv in enumerate(self.lateral_convs)
+        ]
+        # build P6-P7 on top of P5
+        for downsample in self.extra_downsamples:
+            feats.append(downsample(feats[-1]))
+
+        p3, p4, p5, p6, p7 = feats
+
+        for stage in self.fpn_stages:
+            # gp(p6, p4) -> p4_1
+            p4_1 = stage['gp_64_4'](p6, p4, out_size=p4.shape[-2:])
+            # sum(p4_1, p4) -> p4_2
+            p4_2 = stage['sum_44_4'](p4_1, p4, out_size=p4.shape[-2:])
+            # sum(p4_2, p3) -> p3_out
+            p3 = stage['sum_43_3'](p4_2, p3, out_size=p3.shape[-2:])
+            # sum(p3_out, p4_2) -> p4_out
+            p4 = stage['sum_34_4'](p3, p4_2, out_size=p4.shape[-2:])
+            # sum(p5, gp(p4_out, p3_out)) -> p5_out
+            p5_tmp = stage['gp_43_5'](p4, p3, out_size=p5.shape[-2:])
+            p5 = stage['sum_55_5'](p5, p5_tmp, out_size=p5.shape[-2:])
+            # sum(p7, gp(p5_out, p4_2)) -> p7_out
+            p7_tmp = stage['gp_54_7'](p5, p4_2, out_size=p7.shape[-2:])
+            p7 = stage['sum_77_7'](p7, p7_tmp, out_size=p7.shape[-2:])
+            # gp(p7_out, p5_out) -> p6_out
+            p6 = stage['gp_75_6'](p7, p5, out_size=p6.shape[-2:])
+
+        return p3, p4, p5, p6, p7

PyTorch/NLP/Conformer-main/mmdetection/mmdet/models/necks/nasfcos_fpn.py

+import torch.nn as nn
+import torch.nn.functional as F
+from mmcv.cnn import ConvModule, caffe2_xavier_init
+from mmcv.ops.merge_cells import ConcatCell
+
+from ..builder import NECKS
+
+
+@NECKS.register_module()
+class NASFCOS_FPN(nn.Module):
+    """FPN structure in NASFPN.
+
+    Implementation of paper `NAS-FCOS: Fast Neural Architecture Search for
+    Object Detection <https://arxiv.org/abs/1906.04423>`_
+
+    Args:
+        in_channels (List[int]): Number of input channels per scale.
+        out_channels (int): Number of output channels (used at each scale)
+        num_outs (int): Number of output scales.
+        start_level (int): Index of the start input backbone level used to
+            build the feature pyramid. Default: 0.
+        end_level (int): Index of the end input backbone level (exclusive) to
+            build the feature pyramid. Default: -1, which means the last level.
+        add_extra_convs (bool): It decides whether to add conv
+            layers on top of the original feature maps. Default to False.
+            If True, its actual mode is specified by `extra_convs_on_inputs`.
+        conv_cfg (dict): dictionary to construct and config conv layer.
+        norm_cfg (dict): dictionary to construct and config norm layer.
+    """
+
+    def __init__(self,
+                 in_channels,
+                 out_channels,
+                 num_outs,
+                 start_level=1,
+                 end_level=-1,
+                 add_extra_convs=False,
+                 conv_cfg=None,
+                 norm_cfg=None):
+        super(NASFCOS_FPN, self).__init__()
+        assert isinstance(in_channels, list)
+        self.in_channels = in_channels
+        self.out_channels = out_channels
+        self.num_ins = len(in_channels)
+        self.num_outs = num_outs
+        self.norm_cfg = norm_cfg
+        self.conv_cfg = conv_cfg
+
+        if end_level == -1:
+            self.backbone_end_level = self.num_ins
+            assert num_outs >= self.num_ins - start_level
+        else:
+            self.backbone_end_level = end_level
+            assert end_level <= len(in_channels)
+            assert num_outs == end_level - start_level
+        self.start_level = start_level
+        self.end_level = end_level
+        self.add_extra_convs = add_extra_convs
+
+        self.adapt_convs = nn.ModuleList()
+        for i in range(self.start_level, self.backbone_end_level):
+            adapt_conv = ConvModule(
+                in_channels[i],
+                out_channels,
+                1,
+                stride=1,
+                padding=0,
+                bias=False,
+                norm_cfg=dict(type='BN'),
+                act_cfg=dict(type='ReLU', inplace=False))
+            self.adapt_convs.append(adapt_conv)
+
+        # C2 is omitted according to the paper
+        extra_levels = num_outs - self.backbone_end_level + self.start_level
+
+        def build_concat_cell(with_input1_conv, with_input2_conv):
+            cell_conv_cfg = dict(
+                kernel_size=1, padding=0, bias=False, groups=out_channels)
+            return ConcatCell(
+                in_channels=out_channels,
+                out_channels=out_channels,
+                with_out_conv=True,
+                out_conv_cfg=cell_conv_cfg,
+                out_norm_cfg=dict(type='BN'),
+                out_conv_order=('norm', 'act', 'conv'),
+                with_input1_conv=with_input1_conv,
+                with_input2_conv=with_input2_conv,
+                input_conv_cfg=conv_cfg,
+                input_norm_cfg=norm_cfg,
+                upsample_mode='nearest')
+
+        # Denote c3=f0, c4=f1, c5=f2 for convince
+        self.fpn = nn.ModuleDict()
+        self.fpn['c22_1'] = build_concat_cell(True, True)
+        self.fpn['c22_2'] = build_concat_cell(True, True)
+        self.fpn['c32'] = build_concat_cell(True, False)
+        self.fpn['c02'] = build_concat_cell(True, False)
+        self.fpn['c42'] = build_concat_cell(True, True)
+        self.fpn['c36'] = build_concat_cell(True, True)
+        self.fpn['c61'] = build_concat_cell(True, True)  # f9
+        self.extra_downsamples = nn.ModuleList()
+        for i in range(extra_levels):
+            extra_act_cfg = None if i == 0 \
+                else dict(type='ReLU', inplace=False)
+            self.extra_downsamples.append(
+                ConvModule(
+                    out_channels,
+                    out_channels,
+                    3,
+                    stride=2,
+                    padding=1,
+                    act_cfg=extra_act_cfg,
+                    order=('act', 'norm', 'conv')))
+
+    def forward(self, inputs):
+        """Forward function."""
+        feats = [
+            adapt_conv(inputs[i + self.start_level])
+            for i, adapt_conv in enumerate(self.adapt_convs)
+        ]
+
+        for (i, module_name) in enumerate(self.fpn):
+            idx_1, idx_2 = int(module_name[1]), int(module_name[2])
+            res = self.fpn[module_name](feats[idx_1], feats[idx_2])
+            feats.append(res)
+
+        ret = []
+        for (idx, input_idx) in zip([9, 8, 7], [1, 2, 3]):  # add P3, P4, P5
+            feats1, feats2 = feats[idx], feats[5]
+            feats2_resize = F.interpolate(
+                feats2,
+                size=feats1.size()[2:],
+                mode='bilinear',
+                align_corners=False)
+
+            feats_sum = feats1 + feats2_resize
+            ret.append(
+                F.interpolate(
+                    feats_sum,
+                    size=inputs[input_idx].size()[2:],
+                    mode='bilinear',
+                    align_corners=False))
+
+        for submodule in self.extra_downsamples:
+            ret.append(submodule(ret[-1]))
+
+        return tuple(ret)
+
+    def init_weights(self):
+        """Initialize the weights of module."""
+        for module in self.fpn.values():
+            if hasattr(module, 'conv_out'):
+                caffe2_xavier_init(module.out_conv.conv)
+
+        for modules in [
+                self.adapt_convs.modules(),
+                self.extra_downsamples.modules()
+        ]:
+            for module in modules:
+                if isinstance(module, nn.Conv2d):
+                    caffe2_xavier_init(module)