Caffe to keras model AttributeError: type object 'h5py.h5.H5PYConfig' has no attribute '__reduce_cython__'

[alecda573]

Platform : Windows 10

Python version: 3.7.9

Source framework with version: Caffe 1.x Destination framework with version: keras (tensorflow 1.15) Pre-trained model path: OpenPose Body 25 https://github.com/CMU-Perceptual-Computing-Lab/openpose

Running scripts:

I have installed this repo through PyPi using pip install mmdnn

I am trying to convert the model using the following command: mmconvert --srcFramework caffe --inputWeight pose_iter_584000.caffemodel --inputNetwork pose_deploy.prototxt --dstFramework keras --outputModel base.h5

However I get the following error:

File "c:\users\alecd\anaconda3\lib\runpy.py", line 194, in _run_module_as_main
  return _run_code(code, main_globals, None,
File "c:\users\alecd\anaconda3\lib\runpy.py", line 87, in _run_code
  exec(code, run_globals)
File "C:\Users\alecd\anaconda3\Scripts\mmconvert.exe\__main__.py", line 7, in <module>
File "c:\users\alecd\anaconda3\lib\site-packages\mmdnn\conversion\_script\convert.py", line 108, in _main
  ret = IRToCode._convert(code_args)
File "c:\users\alecd\anaconda3\lib\site-packages\mmdnn\conversion\_script\IRToCode.py", line 17, in _convert
  from mmdnn.conversion.keras.keras2_emitter import Keras2Emitter
File "c:\users\alecd\anaconda3\lib\site-packages\mmdnn\conversion\keras\keras2_emitter.py", line 14, in <module>
  from mmdnn.conversion.keras.extra_layers import Scale
File "c:\users\alecd\anaconda3\lib\site-packages\mmdnn\conversion\keras\extra_layers.py", line 8, in <module>
  from keras.engine import Layer, InputSpec
File "C:\Users\alecd\AppData\Roaming\Python\Python38\site-packages\keras\__init__.py", line 3, in <module>
  from . import utils
File "C:\Users\alecd\AppData\Roaming\Python\Python38\site-packages\keras\utils\__init__.py", line 6, in <module>
  from . import conv_utils
File "C:\Users\alecd\AppData\Roaming\Python\Python38\site-packages\keras\utils\conv_utils.py", line 9, in <module>
  from .. import backend as K
File "C:\Users\alecd\AppData\Roaming\Python\Python38\site-packages\keras\backend\__init__.py", line 89, in <module>
  from .tensorflow_backend import *
File "C:\Users\alecd\AppData\Roaming\Python\Python38\site-packages\keras\backend\tensorflow_backend.py", line 5, in <module>
  import tensorflow as tf
File "c:\users\alecd\anaconda3\lib\site-packages\tensorflow\__init__.py", line 41, in <module>
  from tensorflow.python.tools import module_util as _module_util
File "c:\users\alecd\anaconda3\lib\site-packages\tensorflow\python\__init__.py", line 84, in <module>
  from tensorflow.python import keras
File "c:\users\alecd\anaconda3\lib\site-packages\tensorflow\python\keras\__init__.py", line 27, in <module>
  from tensorflow.python.keras import models
File "c:\users\alecd\anaconda3\lib\site-packages\tensorflow\python\keras\models.py", line 24, in <module>
  from tensorflow.python.keras import metrics as metrics_module
File "c:\users\alecd\anaconda3\lib\site-packages\tensorflow\python\keras\metrics.py", line 37, in <module>
  from tensorflow.python.keras.engine import base_layer
File "c:\users\alecd\anaconda3\lib\site-packages\tensorflow\python\keras\engine\base_layer.py", line 59, in <module>
  from tensorflow.python.keras.saving.saved_model import layer_serialization
File "c:\users\alecd\anaconda3\lib\site-packages\tensorflow\python\keras\saving\saved_model\layer_serialization.py", line 24, in <module>
  from tensorflow.python.keras.saving.saved_model import save_impl
File "c:\users\alecd\anaconda3\lib\site-packages\tensorflow\python\keras\saving\saved_model\save_impl.py", line 34, in <module>
  from tensorflow.python.keras.saving import saving_utils
File "c:\users\alecd\anaconda3\lib\site-packages\tensorflow\python\keras\saving\saving_utils.py", line 31, in <module>
  from tensorflow.python.keras.utils.io_utils import ask_to_proceed_with_overwrite
File "c:\users\alecd\anaconda3\lib\site-packages\tensorflow\python\keras\utils\io_utils.py", line 31, in <module>
  import h5py
File "c:\users\alecd\anaconda3\lib\site-packages\h5py\__init__.py", line 34, in <module>
  from . import version
File "c:\users\alecd\anaconda3\lib\site-packages\h5py\version.py", line 17, in <module>
  from . import h5 as _h5
File "h5py\h5.pyx", line 41, in init h5py.h5
AttributeError: type object 'h5py.h5.H5PYConfig' has no attribute '__reduce_cython__'

I have 2.8 of h5py in my environment

I can remove this error by installing keras 2.3.1 directly in my environment but then it throws the following error:

AttributeError: module 'tensorflow' has no attribute 'placeholder' which is a TF v2.x issue, and can only be bypassed to my knowledge by import tensorflow.compat.v1 and disabling v2 behavior. However, I would rather not have to dig through the source if there is another obvious fix.

I successfully can get the intermediate representation, and I have successfully converted the model to onnx but when I check the ``` onnx model using onnx.checker.check I get the following bad node: ValidationError: Node (prelu4_2) has input size 1 not in range [min=2, max=2].

==> Context: Bad node spec: conv4_2prelu4_2prelu4_2"PRelu* slope

[cookieli]

Thank you for your issures, now mmdnn only support tensorflow1.x . And maybe you have wrong tf version

[alecda573]

@cookieli my environment had tensorflow1.15 installed and this did not fix the issue....