Skip to content

Conversation

@HydrogenSulfate
Copy link
Contributor

PR types

Function optimization

PR changes

APIs

Describe

  1. 更新cross_entropy边界检查时的报错代码,让输出的错误信息更清晰
  • label超出上界的case

    import paddle
    logits = paddle.uniform(shape=[1, 10])
    label = paddle.to_tensor([100], dtype='int64')
    loss = paddle.nn.functional.cross_entropy(input=logits, label=label)
    
    # ---输出---
    Traceback (most recent call last):
      File "xxx/ce_loss_optimize.py", line 31, in <module>
        loss = paddle.nn.functional.cross_entropy(input=logits, label=label)
      File "xxx/python3.7/site-packages/paddle/nn/functional/loss.py", line 1662, in cross_entropy
        format(label_max.item()))
    ValueError: Target 100 is out of upper bound.
  • label超出下界的case

    import paddle
    
    logits = paddle.uniform(shape=[1, 10])
    label = paddle.to_tensor([-2], dtype='int64')
    loss = paddle.nn.functional.cross_entropy(input=logits, label=label)
    
    # ---输出---
    Traceback (most recent call last):
      File "xxx/ce_loss_optimize.py", line 35, in <module>
        loss = paddle.nn.functional.cross_entropy(input=logits, label=label)
      File "xxx/python3.7/site-packages/paddle/nn/functional/loss.py", line 1659, in cross_entropy
        format(label_min.item()))
    ValueError: Target -2 is out of lower bound.

@paddle-bot-old
Copy link

Thanks for your contribution!
Please wait for the result of CI firstly. See Paddle CI Manual for details.

Copy link
Contributor

@chajchaj chajchaj left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

@XiaoguangHu01 XiaoguangHu01 merged commit a7acfc5 into PaddlePaddle:develop Mar 1, 2022
@HydrogenSulfate HydrogenSulfate deleted the fix_ce_output branch March 2, 2022 03:10
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants