add nll loss API #26019
Merged
add nll loss API #26019
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
Thanks for your contribution! |
joey12300
force-pushed
the
add_nll_loss
branch
3 times, most recently
from
693432b
to
60e7df1
Compare
wawltor
reviewed
Aug 7, 2020
python/paddle/nn/functional/loss.py
Outdated
| if x_dims != 2 and x_dims != 4: | ||
| x = paddle.reshape(x, shape=[n, c, 1, -1]) | ||
| label = paddle.reshape(label, shape=[n, 1, -1]) | ||
| out_shape = [n] + x_shape[2:] |
python/paddle/nn/functional/loss.py
Outdated
| ignore_index, 'reduction', | ||
| reduction) | ||
| if x_dims != 2 and x_dims != 4 and reduction == 'none': | ||
| out = paddle.reshape(out, shape=out_shape) |
python/paddle/nn/layer/loss.py
Outdated
| @@ -505,18 +506,18 @@ class NLLLoss(fluid.dygraph.Layer): | |||
| \\end{cases} | |||
|
|
|||
| Parameters: | |||
| input (Variable): Input tensor, the data type is float32, float64. | |||
| x (Variable): Input tensor, the data type is float32, float64. | |||
There was a problem hiding this comment.
Variable -> Tensor 下面都是把Variable改成Tensor
wawltor
reviewed
Aug 7, 2020
python/paddle/nn/functional/loss.py
Outdated
| raise ValueError( | ||
| "The value of 'reduction' in nll_loss should be 'sum', 'mean' or " | ||
| "'none', but received %s, which is not allowed." % reduction) | ||
| x_shape = list(x.shape) |
joey12300
force-pushed
the
add_nll_loss
branch
3 times, most recently
from
826b71b
to
ee9ee56
Compare
joey12300
force-pushed
the
add_nll_loss
branch
4 times, most recently
from
f659ddf
to
493292a
Compare
wawltor
previously approved these changes
Aug 10, 2020
python/paddle/nn/functional/loss.py
Outdated
|
|
||
| Parameters: | ||
| x (Tensor): Input tensor, the data type is float32, float64. | ||
| label (Tensor): Label tensor, the data type is int64_t. |
python/paddle/nn/functional/loss.py
Outdated
| and does not contribute to the input gradient. | ||
| reduction (str, optional): Indicate how to average the loss, | ||
| the candicates are ``'none'`` | ``'mean'`` | ``'sum'``. | ||
| If :attr:`reduction` is ``'mean'``, the reduced mean loss is returned; |
python/paddle/nn/functional/loss.py
Outdated
| prog, | ||
| feed={"x": x_np, | ||
| "label": label_np}, | ||
| fetch_list=[res]) |
python/paddle/nn/functional/loss.py
Outdated
| label = paddle.reshape(label, shape=[n, 1, -1]) | ||
| out_shape = [n] + x_shape[2:] | ||
|
|
||
| helper = LayerHelper('nll_loss', **locals()) |
| reduction (str, optional): Indicate how to average the loss, | ||
| the candicates are ``'none'`` | ``'mean'`` | ``'sum'``. | ||
| If :attr:`reduction` is ``'mean'``, the reduced mean loss is returned; | ||
| Default is ``'mean'``. | ||
| ignore_index (int64, optional): Specifies a target value that is ignored | ||
| and does not contribute to the input gradient. |
python/paddle/nn/layer/loss.py
Outdated
|
|
||
| Returns: | ||
| The tensor variable storing the nll_loss. | ||
| The callable object which calculates negative log likelihood loss when | ||
| get the input `x` and `label` . |
joey12300
force-pushed
the
add_nll_loss
branch
from
493292a
to
08cee15
Compare
TCChenlong
reviewed
Aug 12, 2020
python/paddle/nn/functional/loss.py
Outdated
| and does not contribute to the input gradient. | ||
| reduction (str, optional): Indicate how to average the loss, | ||
| the candicates are ``'none'`` | ``'mean'`` | ``'sum'``. | ||
| If :attr:`reduction` is ``'mean'``, the reduced mean loss is returned; |
There was a problem hiding this comment.
格式应该有问题 【If reduction is】就可以 下面同理
zhiqiu
reviewed
Aug 12, 2020
python/paddle/nn/functional/loss.py
Outdated
| "'none', but received %s, which is not allowed." % reduction) | ||
|
|
||
| if in_dygraph_mode(): | ||
| x_shape = list(core.ops.shape(x)) |
There was a problem hiding this comment.
you can write x_shape = x.shape directly here
python/paddle/nn/layer/loss.py
Outdated
| "The value of 'reduction' in nll_loss should be 'sum', 'mean' or " | ||
| "'none', but received %s, which is not allowed." % reduction) | ||
| super(NLLLoss, self).__init__() | ||
| self.weight = weight |
There was a problem hiding this comment.
maybe self._weight is better for private member, etc.
zhiqiu
reviewed
Aug 12, 2020
python/paddle/nn/functional/loss.py
Outdated
| x (Tensor): Input tensor, the data type is float32, float64. | ||
| label (Tensor): Label tensor, the data type is int64. | ||
| weight (Tensor, optional): Weight tensor, a manual rescaling weight given | ||
| to each class. If given, it has to be a Tensor of size `C`. Otherwise, |
There was a problem hiding this comment.
it has to be a Tensor of size
C
Better describe this clearly, does it means x should be of format [N, C, H, W]
There was a problem hiding this comment.
it has to be a Tensor of size
CBetter describe this clearly, does it means x should be of format [N, C, H, W]
@zhiqiu Already describe the shape of x, label, weight more clearly.
joey12300
force-pushed
the
add_nll_loss
branch
2 times, most recently
from
0c5698e
to
349c5e8
Compare
joey12300
force-pushed
the
add_nll_loss
branch
from
349c5e8
to
ede9a87
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
PR types
New features
PR changes
APIs
Describe
Add nll_loss function, and encapsulate NLL_Loss class and optimize the performance of nll_loss in dygraph mode