MMdnn is a set of tools to help users inter-operate among different deep learning frameworks. E.g. model conversion and visualization. Convert models between Caffe, Keras, MXNet, Tensorflow, CNTK, PyTorch Onnx and CoreML.
MIT License
5.8k
stars
965
forks
source link
keras Mobilenetv2 convert to TF/CNTK not support ReLU #554
mmconvert -sf keras -iw MOBT-128-GRAY-acc-0.9792-loss-0.0502.h5 -df tensorflow -om MOBT-128.pb
&
mmconvert -sf keras -iw MOBT-128-GRAY-acc-0.9792-loss-0.0502.h5 -df cntk -om MOBT-128.dnn
Part of Log
KerasParser has not supported operator [ReLU].
{'name': 'Conv1_relu', 'trainable': True, 'max_value': array(6., dtype=float32), 'negative_slope': array(0., dtype=float32), 'threshold': array(0., dtype=float32)}
KerasParser has not supported operator [ReLU].
Platform (like ubuntu 16.04/win10): Win 10
Python version:3.6
Source framework with version (like Tensorflow 1.4.1 with GPU):Keras 2.2.4
Destination framework with version (like CNTK 2.3 with GPU):CNTK 2.6 with GPU
Pre-trained model path (webpath or webdisk path): https://drive.google.com/file/d/1xt6gHkFdt8mePFNZsEu6bq5mdSz6O-Lj/view?usp=sharing
Running scripts:
mmconvert -sf keras -iw MOBT-128-GRAY-acc-0.9792-loss-0.0502.h5 -df tensorflow -om MOBT-128.pb & mmconvert -sf keras -iw MOBT-128-GRAY-acc-0.9792-loss-0.0502.h5 -df cntk -om MOBT-128.dnn