提交 24144bdb 编写于 作者: W wanghaoshuang

Add notice about norm_by_times and change 'ctc' to 'CTC'

上级 fb62f8cb
...@@ -1508,10 +1508,11 @@ def reduce_min(input, dim=None, keep_dim=False): ...@@ -1508,10 +1508,11 @@ def reduce_min(input, dim=None, keep_dim=False):
def warpctc(input, label, blank=0, norm_by_times=False, **kwargs): def warpctc(input, label, blank=0, norm_by_times=False, **kwargs):
""" """
An operator integrating the open source warp-ctc library An operator integrating the open source Warp-CTC library
(https://github.com/baidu-research/warp-ctc)
to compute Connectionist Temporal Classification (CTC) loss. to compute Connectionist Temporal Classification (CTC) loss.
It can be aliased as softmax with ctc, since a native softmax activation is It can be aliased as softmax with CTC, since a native softmax activation is
interated to the warp-ctc library, to to normlize values for each row of the interated to the Warp-CTC library, to to normlize values for each row of the
input tensor. input tensor.
Args: Args:
...@@ -1525,12 +1526,12 @@ def warpctc(input, label, blank=0, norm_by_times=False, **kwargs): ...@@ -1525,12 +1526,12 @@ def warpctc(input, label, blank=0, norm_by_times=False, **kwargs):
of variable-length sequence, which is a 2-D Tensor with LoD of variable-length sequence, which is a 2-D Tensor with LoD
information. It is of the shape [Lg, 1], where Lg is th sum of information. It is of the shape [Lg, 1], where Lg is th sum of
all labels' length. all labels' length.
blank: (int, default: 0), the blank label of Connectionist blank: (int, default: 0), the blank label index of Connectionist
Temporal Classification (CTC) loss, which is in the Temporal Classification (CTC) loss, which is in the
half-opened interval [0, num_classes + 1). half-opened interval [0, num_classes + 1).
norm_by_times: (bool, default: false), whether to norm_by_times: (bool, default: false), whether to normalize the gradients
normalize the gradients by the number of time-step, by the number of time-step,which is also the sequence's length.
which is also the sequence's length. There is no need to normalize the gradients if warpctc layer was follewed by a mean_op.
Returns: Returns:
Variable: The Connectionist Temporal Classification (CTC) loss, which is a 2-D Tensor of the shape [batch_size, 1]. Variable: The Connectionist Temporal Classification (CTC) loss, which is a 2-D Tensor of the shape [batch_size, 1].
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册