Created by: xubin1994

我使用 unsqueeze 的示例程序能正确扩展维度： x = fluid.layers.data(name='x', shape=[5, 10]) y = fluid.layers.unsqueeze(input=x, axes=[1]) 但是当我使用自己读取的图片数据，在动态图中转变为 variable 之后，却总是报错： with fluid.dygraph.guard(): img_l = fluid.dygraph.to_variable(img_l) img_r = fluid.dygraph.to_variable(img_r) disp_l = fluid.dygraph.to_variable(disp_l) disp_r = fluid.dygraph.to_variable(disp_r) img_l = paddle.fluid.layers.stack([img_l],axis=0) img_r = paddle.fluid.layers.stack([img_r],axis=0) disp = paddle.fluid.layers.unsqueeze(input=disp_l, axes=[0]) ··· 报错信息如下： D:\ProgramData\Anaconda3\envs\deeplearning\python.exe D:/IresNet-PaddlePaddle/stn.py Traceback (most recent call last): File "D:/IresNet-PaddlePaddle/stn.py", line 188, in disp = paddle.fluid.layers.unsqueeze(input=disp_l, axes=[0]) File "D:\ProgramData\Anaconda3\envs\deeplearning\lib\site-packages\paddle\fluid\layers\nn.py", line 6941, in unsqueeze "XShape": x_shape}) File "D:\ProgramData\Anaconda3\envs\deeplearning\lib\site-packages\paddle\fluid\layer_helper.py", line 43, in append_op return self.main_program.current_block().append_op(*args, **kwargs) File "D:\ProgramData\Anaconda3\envs\deeplearning\lib\site-packages\paddle\fluid\framework.py", line 1762, in append_op kwargs.get("stop_gradient", False)) File "D:\ProgramData\Anaconda3\envs\deeplearning\lib\site-packages\paddle\fluid\dygraph\tracer.py", line 59, in trace_op framework._current_expected_place(), stop_gradient) paddle.fluid.core_avx.EnforceNotMet: op_kernel should not be null only support op with kernel at [D:\1.5.1\release_cuda97\paddle\paddle\fluid\imperative\tracer.cc:217] PaddlePaddle Call Stacks: Windows not support stack backtrace yet.

Process finished with exit code 1 请问是什么问题呢，我的图片数据是读取正确的，并转换为 numpy 数组