dice loss #396
dice loss #396
Conversation
|
|
Codecov Report
@@ Coverage Diff @@
## master #396 +/- ##
==========================================
+ Coverage 86.20% 86.32% +0.12%
==========================================
Files 96 98 +2
Lines 4906 4973 +67
Branches 799 808 +9
==========================================
+ Hits 4229 4293 +64
- Misses 523 525 +2
- Partials 154 155 +1
Flags with carried forward coverage won't be shown. Click here to find out more.
Continue to review full report at Codecov.
|
|
Hi @xiexinch |
mmseg/models/losses/dice_loss.py
Outdated
| class DiceLoss(nn.Module): | ||
| """DiceLoss. | ||
|
|
||
| """ |
There was a problem hiding this comment.
We may add some docstring here.
mmseg/models/losses/dice_loss.py
Outdated
| valid_mask = valid_mask.contiguous().view(valid_mask.shape[0], -1) | ||
|
|
||
| num = torch.sum(torch.mul(pred, target) * valid_mask, dim=1) * 2 + smooth | ||
| den = torch.sum((pred.pow(exponent) + target.pow(exponent)) * valid_mask, dim=1) + smooth |
There was a problem hiding this comment.
We may directly use denominator.
mmseg/models/losses/dice_loss.py
Outdated
| class_weight (list[float], optional): The weight for each class. | ||
| Default: None. | ||
| loss_weight (float, optional): Weight of the loss. Default to 1.0. | ||
| ignore_index (int | None): The label index to be ignored. Default: -1. |
|
FCN dilate model shouldn't be in this PR |
| else: | ||
| class_weight = None | ||
|
|
||
| pred = F.softmax(pred, dim=1) |
There was a problem hiding this comment.
In instance segmentation, dice loss may use sigmoid for activation. Suggest supporting both cases.
| assert pred.shape[0] == target.shape[0] | ||
| total_loss = 0 | ||
| num_classes = pred.shape[1] | ||
| for i in range(num_classes): |
There was a problem hiding this comment.
Use for loop might be inefficient? Some implementation support to process multi-class in a batched manner.
There was a problem hiding this comment.
| @weighted_loss | ||
| def binary_dice_loss(pred, target, valid_mask, smooth=1, exponent=2, **kwards): | ||
| assert pred.shape[0] == target.shape[0] | ||
| pred = pred.contiguous().view(pred.shape[0], -1) |
There was a problem hiding this comment.
.contiguous().view() can be replaced by reshape?
* dice loss * format code, add docstring and calculate denominator without valid_mask * minor change * restore
* ddim docs for issue open-mmlab#293 * space
No description provided.