Error from `avg_pool3d` & `Conv3D ` in 3D_CNN

[Le0v1n]

非常棒的Repo，但我想使用convert.py将PyTorch训练好的3DCNN分类模型转换为.tflite模型，发生了错误：

问题1：x = F.avg_pool3d(x, x.data.size()[-3:])不支持

ERROR (tinynn.converter.base) Unsupported ops: aten::avg_pool3d

问题2：我将Temporal通道Squeeze掉了，但还是发生了错误：

Traceback (most recent call last):
  File "/home/leovin/Dynamic Gesture Detection/3D-CNNs/TinyNeuralNetwork/examples/converter/convert_for_3DCNNs.py", line 50, in <module>
    main_worker()
  File "/home/leovin/Dynamic Gesture Detection/3D-CNNs/TinyNeuralNetwork/examples/converter/convert_for_3DCNNs.py", line 46, in main_worker
    converter.convert()
  File "/home/leovin/anaconda3/envs/dgr/lib/python3.9/site-packages/TinyNeuralNetwork-0.1.20220310142810-py3.9.egg/tinynn/converter/base.py", line 347, in convert
    optimizer.optimize()
  File "/home/leovin/anaconda3/envs/dgr/lib/python3.9/site-packages/TinyNeuralNetwork-0.1.20220310142810-py3.9.egg/tinynn/converter/operators/optimize.py", line 1377, in optimize
    self.transform_graph()
  File "/home/leovin/anaconda3/envs/dgr/lib/python3.9/site-packages/TinyNeuralNetwork-0.1.20220310142810-py3.9.egg/tinynn/converter/operators/optimize.py", line 210, in transform_graph
    op.transform(self.graph, mapping)
  File "/home/leovin/anaconda3/envs/dgr/lib/python3.9/site-packages/TinyNeuralNetwork-0.1.20220310142810-py3.9.egg/tinynn/converter/operators/tflite/transformable.py", line 235, in transform
    assert False, "Only Conv[Transpose]1d/2d is supported"
AssertionError: Only Conv[Transpose]1d/2d is supported

Process finished with exit code 1

Only Conv[Transpose]1d/2d is supported请问是因为不支持Conv3d吗？

问题3：如果支持Conv3d，我该怎么做？

期待您的回答！

[peterjc123]

问题1： avg_pool3d 目前不支持，新版本支持 问题2： 是的，目前不支持conv3d 问题3：你是问要怎么支持conv3d吗？

[Le0v1n]

我想将我的3D-CNN分类模型运行在手机上，所以如何将含有Conv3d的PyTorch模型转换为.tflite模型呢？

[peterjc123]

需要converter这边支持conv3d的转换，但是麻烦的是这个会牵扯到tflite schema的升级，因为2.4和2.3完全不兼容

[Le0v1n]

需要converter这边支持conv3d的转换，但是麻烦的是这个会牵扯到tflite schema的升级，因为2.4和2.3完全不兼容

意思是这个问题占时无解，只能等待后续支持Conv3d了，是吗？

请问还有其他办法来实现同样的功能吗？

[peterjc123]

除非你kernel的维度有1，可以降成2D来跑，否则好像是不行，只能等升级了

[Le0v1n]

除非你kernel的维度有1，可以降成2D来跑，否则好像是不行，只能等升级了

嗯嗯，等升级吧 :joy:

非常感谢你的回答！

[peterjc123]

@Le0v1n 抱歉，纠正一下，看这里https://github.com/tensorflow/tensorflow/blob/master/tensorflow/lite/kernels/pooling3d.cc 3d pooling在新版本TFLite中是支持的

[Le0v1n]

收到

[peterjc123]

相关问题：https://github.com/alibaba/TinyNeuralNetwork/issues/50

[peterjc123]

TFLite schema升级相关PR: https://github.com/alibaba/TinyNeuralNetwork/pull/57

[peterjc123]

@Le0v1n 3D卷积和反卷积已经支持了 https://github.com/alibaba/TinyNeuralNetwork/commit/468801a4179506109bebbceb67e7b13af11020bb

[Le0v1n]

@Le0v1n 3D卷积和反卷积已经支持了 468801a

谢谢大佬！大佬辛苦了！

[peterjc123]

@Le0v1n 帮忙试下模型现在可以转换了不？

[Le0v1n]

@Le0v1n 帮忙试下模型现在可以转换了不？

Some ops are not supported by the native TFLite runtime, you can enable TF kernels fallback using TF Select. See instructions: https://www.tensorflow.org/lite/guide/ops_select 
TF Select ops: AvgPool3D
Details:
    tf.AvgPool3D(tensor<1x1x6x6x1024xf32>) -> (tensor<1x1x1x1x1024xf32>) : {data_format = "NDHWC", device = "", ksize = [1, 1, 6, 6, 1], padding = "VALID", strides = [1, 1, 6, 6, 1]}

nn.AvgPool3d仍然不支持，但3D-CNN可以了。

非常感谢大佬！大佬辛苦了！