search

microsoft / MMdnn

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
5.79k stars 964 forks source link

tensorflow official tutorial model to caffe #692

Open allendred opened 5 years ago

allendred commented 5 years ago

Platform (like ubuntu 16.04/win10): 16.04 Python version: 3.5 Source framework with version (like Tensorflow 1.4.1 with GPU): 1.8 whl Destination framework with version (like CNTK 2.3 with GPU): caffe gpu Pre-trained model path (webpath or webdisk path):

Running scripts:

I0628 15:52:33.465559 11669 net.cpp:122] Setting up resnet_v2_152_block3_unit_11_bottleneck_v2_preact_Relu I0628 15:52:33.465572 11669 net.cpp:129] Top shape: 1 1024 19 19 (369664) I0628 15:52:33.465576 11669 net.cpp:137] Memory required for data: 491605164 I0628 15:52:33.465582 11669 layer_factory.hpp:77] Creating layer resnet_v2_152_block3_unit_11_bottleneck_v2_conv1_Conv2D I0628 15:52:33.465590 11669 net.cpp:84] Creating Layer resnet_v2_152_block3_unit_11_bottleneck_v2_conv1_Conv2D I0628 15:52:33.465596 11669 net.cpp:406] resnet_v2_152_block3_unit_11_bottleneck_v2_conv1_Conv2D <- resnet_v2_152_block3_unit_11_bottleneck_v2_preact_FusedBatchNorm I0628 15:52:33.465602 11669 net.cpp:380] resnet_v2_152_block3_unit_11_bottleneck_v2_conv1_Conv2D -> resnet_v2_152_block3_unit_11_bottleneck_v2_conv1_Conv2D

F0628 15:52:33.467706 11669 cudnn_conv_layer.cpp:53] Check failed: status == CUDNN_STATUS_SUCCESS (2 vs. 0) CUDNN_STATUS_ALLOC_FAILED

Check failure stack trace:

allendred commented 5 years ago

After pip uninstall cudnn,still wrong detail: F0628 16:13:27.567560 12096 cudnn_conv_layer.cpp:53] Check failed: status == CUDNN_STATUS_SUCCESS (1 vs. 0) CUDNN_STATUS_NOT_INITIALIZED Check failure stack trace: