CrossEntropyWithSoftmaxCUDAKernel

[cccrice]

欢迎您反馈PaddleSeg使用问题，辛苦您提供以下信息，方便我们快速定位和解决问题：

1. PaddleSeg版本：无
2. PaddlePaddle版本：PaddlePaddle 2.3.0
3. 操作系统信息：Windows10
4. Python版本号：Python3.6
5. CUDA/cuDNN版本：CUDA10.2/cuDNN 7.6.5
6. 完整的代码：(若修改过原代码，请提供修改前后代码对比）
7. 详细的错误信息、相关log：Traceback (most recent call last): File "G:/学习/飞桨学习课程/语义分割/case_pet_seg/Unet.py", line 163, in batch_size=2) File "C:\Program Files\Python36\lib\site-packages\paddle\hapi\model.py", line 1767, in fit logs = self._run_one_epoch(train_loader, cbks, 'train') File "C:\Program Files\Python36\lib\site-packages\paddle\hapi\model.py", line 2097, in _run_one_epoch outs = getattr(self, mode + '_batch')(_inputs) File "C:\Program Files\Python36\lib\site-packages\paddle\hapi\model.py", line 1093, in train_batch loss = self._adapter.train_batch(inputs, labels, update) File "C:\Program Files\Python36\lib\site-packages\paddle\hapi\model.py", line 730, in train_batch losses = self.model._loss((to_list(outputs) + labels)) File "C:\Program Files\Python36\lib\site-packages\paddle\fluid\dygraph\layers.py", line 930, in call return self._dygraph_call_func(*inputs, **kwargs) File "C:\Program Files\Python36\lib\site-packages\paddle\fluid\dygraph\layers.py", line 915, in _dygraph_call_func outputs = self.forward(*inputs, **kwargs) File "C:\Program Files\Python36\lib\site-packages\paddle\nn\layer\loss.py", line 406, in forward name=self.name) File "C:\Program Files\Python36\lib\site-packages\paddle\nn\functional\loss.py", line 1734, in cross_entropy 'use_softmax', use_softmax) SystemError: (Fatal) Operator softmax_with_cross_entropy raises an struct paddle::PD_Exception exception. The exception content is :function "CrossEntropyWithSoftmaxCUDAKernel" is not implemented for data type float32 [../paddle/phi/kernels/gpu/cross_entropy_kernel.cu:1544]. (at ..\paddle\fluid\imperative\tracer.cc:307)
8. 运行指令或复现步骤：
9. 其他内容: （增加其他与问题相关的内容）

[juncaipeng]

你看下使用的是那个loss api，参考官方api的说明检查一下输入类型是否匹配。