def forward(self, outputs, targets):
""" Performs the matching
Params:
outputs: This is a dict that contains at least these entries:
"pred_logits": Tensor of dim [batch_size, num_queries, num_classes] with the
classification logits
"pred_boxes": Tensor of dim [batch_size, num_queries, 4] with the predicted
box coordinates
targets: This is a list of targets (len(targets) = batch_size), where each target
is a dict containing:
"labels": Tensor of dim [num_target_boxes] (where num_target_boxes is the number
of ground-truth objects in the target) containing the class labels
"boxes": Tensor of dim [num_target_boxes, 4] containing the target box coordinates
Returns:
A list of size batch_size, containing tuples of (index_i, index_j) where:
- index_i is the indices of the selected predictions (in order)
- index_j is the indices of the corresponding selected targets (in order)
For each batch element, it holds:
len(index_i) = len(index_j) = min(num_queries, num_target_boxes)
"""
bs, num_queries = outputs["pred_logits"].shape[:2]
# We flatten to compute the cost matrices in a batch
out_prob = (outputs["pred_logits"].flatten(0, 1).softmax(-1)
) # [batch_size * num_queries, num_classes]
out_bbox = outputs["pred_boxes"].flatten(
0, 1) # [batch_size * num_queries, 4]
# Also concat the target labels and boxes
tgt_ids = torch.cat([v["labels"] for v in targets])
tgt_bbox = torch.cat([v["boxes"] for v in targets])
# Compute the classification cost. Contrary to the loss, we don't use the NLL,
# but approximate it in 1 - proba[target class].
# The 1 is a constant that doesn't change the matching, it can be ommitted.
cost_class = -out_prob[:, tgt_ids]
# Compute the L1 cost between boxes
cost_bbox = torch.cdist(out_bbox, tgt_bbox, p=1)
# Compute the giou cost betwen boxes
cost_giou = -generalized_box_iou(box_cxcywh_to_xyxy(out_bbox),
box_cxcywh_to_xyxy(tgt_bbox))
# Final cost matrix
C = self.cost_bbox * cost_bbox + self.cost_class * cost_class + self.cost_giou * cost_giou
C = C.view(bs, num_queries, -1).cpu()
sizes = [len(v["boxes"]) for v in targets]
indices = [
linear_sum_assignment(c[i])
for i, c in enumerate(C.split(sizes, -1))
]
return [(torch.as_tensor(i, dtype=torch.int64),
torch.as_tensor(j, dtype=torch.int64)) for i, j in indices]
def loss_boxes(self, outputs, targets, indices, num_boxes):
"""
Compute the losses related to the bounding boxes, the L1 regression loss and the GIoU loss
targets dicts must contain the key "boxes" containing a tensor of dim [nb_target_boxes, 4]
The target boxes are expected in format (center_x, center_y, h, w), normalized by the
image size.
"""
assert "pred_boxes" in outputs
idx = self._get_src_permutation_idx(indices)
src_boxes = outputs["pred_boxes"][idx]
target_boxes = torch.cat(
[t["boxes"][i] for t, (_, i) in zip(targets, indices)], dim=0)
loss_bbox = F.l1_loss(src_boxes, target_boxes, reduction="none")
losses = {}
losses["loss_bbox"] = loss_bbox.sum() / num_boxes
loss_giou = 1 - torch.diag(
generalized_box_iou(box_ops.box_cxcywh_to_xyxy(src_boxes),
box_ops.box_cxcywh_to_xyxy(target_boxes)))
losses["loss_giou"] = loss_giou.sum() / num_boxes
return losses
def forward(self, outputs, targets):
""" Performs the matching
Params:
outputs: This is a dict that contains at least these entries:
"pred_logits": Tensor of dim [batch_size, num_queries, num_classes]
with the classification logits
"pred_boxes": Tensor of dim [batch_size, num_queries, 4]
with the predicted box coordinates
targets: This is a list of targets (len(targets) = batch_size),
where each target is a dict containing:
"labels": Tensor of dim [num_target_boxes]
(where num_target_boxes is the number of ground-truth
objects in the target) containing the class labels
"boxes": Tensor of dim [num_target_boxes, 4]
containing the target box coordinates
Returns:
A list of size batch_size, containing tuples of (index_i, index_j) where:
- index_i is the indices of the selected predictions (in order)
- index_j is the indices of the corresponding selected targets (in order)
For each batch element, it holds:
len(index_i) = len(index_j) = min(num_queries, num_target_boxes)
"""
bs, num_queries = outputs["pred_logits"].shape[:2]
# We flatten to compute the cost matrices in a batch
if self.use_focal:
out_prob = outputs["pred_logits"].flatten(0, 1).sigmoid()
out_bbox = outputs["pred_boxes"].flatten(0, 1)
else:
out_prob = outputs["pred_logits"].flatten(0, 1).softmax(-1)
out_bbox = outputs["pred_boxes"].flatten(0, 1)
# Also concat the target labels and boxes
tgt_ids = torch.cat([v["labels"] for v in targets])
tgt_bbox = torch.cat([v["boxes_xyxy"] for v in targets])
# Compute the classification cost. Contrary to the loss, we don't use the NLL,
# but approximate it in 1 - proba[target class].
# The 1 is a constant that doesn't change the matching, it can be ommitted.
if self.use_focal:
# Compute the classification cost.
alpha = self.focal_loss_alpha
gamma = self.focal_loss_gamma
neg_cost_class = (1 - alpha) * (out_prob ** gamma) * (-(1 - out_prob + 1e-8).log())
pos_cost_class = alpha * ((1 - out_prob) ** gamma) * (-(out_prob + 1e-8).log())
cost_class = pos_cost_class[:, tgt_ids] - neg_cost_class[:, tgt_ids]
else:
cost_class = -out_prob[:, tgt_ids]
# Compute the L1 cost between boxes
image_size_out = torch.cat([v["image_size_xyxy"].unsqueeze(0) for v in targets])
image_size_out = image_size_out.unsqueeze(1).repeat(1, num_queries, 1).flatten(0, 1)
image_size_tgt = torch.cat([v["image_size_xyxy_tgt"] for v in targets])
out_bbox_ = out_bbox / image_size_out
tgt_bbox_ = tgt_bbox / image_size_tgt
cost_bbox = torch.cdist(out_bbox_, tgt_bbox_, p=1)
# Compute the giou cost betwen boxes
cost_giou = -generalized_box_iou(out_bbox, tgt_bbox)
# Final cost matrix
C = self.cost_bbox * cost_bbox + self.cost_class * cost_class + self.cost_giou * cost_giou
C = C.view(bs, num_queries, -1).cpu()
sizes = [len(v["boxes"]) for v in targets]
indices = [linear_sum_assignment(c[i]) for i, c in enumerate(C.split(sizes, -1))]
return [(torch.as_tensor(i, dtype=torch.int64),
torch.as_tensor(j, dtype=torch.int64)) for i, j in indices]
