Skip to content

P
Paddle

PaddlePaddle / Paddle
1 年多 前同步成功

通知 2302
Star 20931
Fork 5422
P
Paddle
提交 03299ed4 编写于 作者: J Jacek Czaja
浏览文件
操作

- Fix to linking for GPU builds of softmax inference 

test=develop
上级 07563437
隐藏空白更改
内联 并排
Showing with 4 addition and 0 deletion
paddle/fluid/operators/math/softmax.cu
浏览文件 @ 03299ed4
...@@ -100,8 +100,12 @@ template class SoftmaxGradCUDNNFunctor<platform::float16>; ...@@ -100,8 +100,12 @@ template class SoftmaxGradCUDNNFunctor<platform::float16>;
template class SoftmaxFunctor<platform::CUDADeviceContext, platform::float16, template class SoftmaxFunctor<platform::CUDADeviceContext, platform::float16,
false>; false>;
template class SoftmaxFunctor<platform::CUDADeviceContext, platform::float16,
true>;
template class SoftmaxFunctor<platform::CUDADeviceContext, float, false>; template class SoftmaxFunctor<platform::CUDADeviceContext, float, false>;
template class SoftmaxFunctor<platform::CUDADeviceContext, double, false>; template class SoftmaxFunctor<platform::CUDADeviceContext, double, false>;
template class SoftmaxFunctor<platform::CUDADeviceContext, float, true>;
template class SoftmaxFunctor<platform::CUDADeviceContext, double, true>;
template class SoftmaxGradFunctor<platform::CUDADeviceContext, float>; template class SoftmaxGradFunctor<platform::CUDADeviceContext, float>;
template class SoftmaxGradFunctor<platform::CUDADeviceContext, double>; template class SoftmaxGradFunctor<platform::CUDADeviceContext, double>;
template class SoftmaxGradFunctor<platform::CUDADeviceContext, template class SoftmaxGradFunctor<platform::CUDADeviceContext,
......
Markdown is supported
0% 或 .
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 或