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
5.78k
stars
968
forks
source link
Fail to Convert resnet from Caffe to onnx and Pytorch #932
I am getting below error while converting Resnet Model From caffe to onnx:
I0217 14:18:15.896023 8924 net.cpp:196] Top shape: 1 24 14 14 (4704)
I0217 14:18:15.896029 8924 layer_factory.hpp:77] Creating layer proposal
ImportError: No module named 'rpn'
Traceback (most recent call last):
File "/home/ubuntu/anaconda3/envs/python_3_5/bin/mmconvert", line 8, in
sys.exit(_main())
File "/home/ubuntu/anaconda3/envs/python_3_5/lib/python3.5/site-packages/mmdnn/conversion/_script/convert.py", line 102, in _main
ret = convertToIR._convert(ir_args)
File "/home/ubuntu/anaconda3/envs/python_3_5/lib/python3.5/site-packages/mmdnn/conversion/_script/convertToIR.py", line 16, in _convert
transformer = CaffeTransformer(args.network, args.weights, "tensorflow", inputshape[0], phase = args.caffePhase)
File "/home/ubuntu/anaconda3/envs/python_3_5/lib/python3.5/site-packages/mmdnn/conversion/caffe/transformer.py", line 325, in init
graph = GraphBuilder(def_path, self.input_shape, self.is_train_proto, phase).build()
File "/home/ubuntu/anaconda3/envs/python_3_5/lib/python3.5/site-packages/mmdnn/conversion/caffe/graph.py", line 454, in build
graph.compute_output_shapes(self.model)
File "/home/ubuntu/anaconda3/envs/python_3_5/lib/python3.5/site-packages/mmdnn/conversion/caffe/graph.py", line 263, in compute_output_shapes
net = caffe.Net(tmp_prototxt, caffe.TEST)
SystemError: <Boost.Python.function object at 0x2ccfd50> returned NULL without setting an error
Anyone can please help to figure it out?
Thank you.
Below are system Specifications and command I used to convert:
Platform (like ubuntu 16.04/win10): ubuntu 16.04
Python version: 3.5
Source framework with version (like Tensorflow 1.4.1 with GPU): CAFFE
Destination framework with version (like CNTK 2.3 with GPU): ONNX or PYTORCH
Hi,
I am getting below error while converting Resnet Model From caffe to onnx:
I0217 14:18:15.896023 8924 net.cpp:196] Top shape: 1 24 14 14 (4704) I0217 14:18:15.896029 8924 layer_factory.hpp:77] Creating layer proposal ImportError: No module named 'rpn' Traceback (most recent call last): File "/home/ubuntu/anaconda3/envs/python_3_5/bin/mmconvert", line 8, in
sys.exit(_main())
File "/home/ubuntu/anaconda3/envs/python_3_5/lib/python3.5/site-packages/mmdnn/conversion/_script/convert.py", line 102, in _main
ret = convertToIR._convert(ir_args)
File "/home/ubuntu/anaconda3/envs/python_3_5/lib/python3.5/site-packages/mmdnn/conversion/_script/convertToIR.py", line 16, in _convert
transformer = CaffeTransformer(args.network, args.weights, "tensorflow", inputshape[0], phase = args.caffePhase)
File "/home/ubuntu/anaconda3/envs/python_3_5/lib/python3.5/site-packages/mmdnn/conversion/caffe/transformer.py", line 325, in init
graph = GraphBuilder(def_path, self.input_shape, self.is_train_proto, phase).build()
File "/home/ubuntu/anaconda3/envs/python_3_5/lib/python3.5/site-packages/mmdnn/conversion/caffe/graph.py", line 454, in build
graph.compute_output_shapes(self.model)
File "/home/ubuntu/anaconda3/envs/python_3_5/lib/python3.5/site-packages/mmdnn/conversion/caffe/graph.py", line 263, in compute_output_shapes
net = caffe.Net(tmp_prototxt, caffe.TEST)
SystemError: <Boost.Python.function object at 0x2ccfd50> returned NULL without setting an error
Anyone can please help to figure it out? Thank you.
Below are system Specifications and command I used to convert: Platform (like ubuntu 16.04/win10): ubuntu 16.04
Python version: 3.5
Source framework with version (like Tensorflow 1.4.1 with GPU): CAFFE
Destination framework with version (like CNTK 2.3 with GPU): ONNX or PYTORCH
Running scripts: mmconvert --srcFramework caffe --inputWeight ./AMAR_MODELS/ctd_tloc_iter_91000_bkup_2_11_CW.caffemodel --inputNetwork ./AMAR_MODELS/test_ctd_tloc.prototxt --dstFramework onnx --outputModel ctd_tloc_iter_91000_bkup_2_11_CW.onnx --inputShape 1 3 224 224