Skip to content
This repository has been archived by the owner on Oct 1, 2019. It is now read-only.

Error when start training: Unknown enumeration value of "VIDEO_DATA" for filed "type" #212

Closed
applesleam opened this issue Mar 9, 2017 · 2 comments

Comments

@applesleam
Copy link

Hi, thank you for releasing C3D-v1.1
when I use ./build/tools/caffe train --solver solver.prototxt --weights ../models/conv3d_deepnetA_sport1m_iter_1900000 to finetune my own network, the following error reported:

I0309 10:56:55.158175 19636 solver.cpp:77] Creating training net from train_net file: train.prototxt
[libprotobuf ERROR google/protobuf/text_format.cc:245] Error parsing text-format caffe.NetParameter: 5:3: Unknown enumeration value of "VIDEO_DATA" for field "type".
F0309 10:56:55.158308 19636 upgrade_proto.cpp:88] Check failed: ReadProtoFromTextFile(param_file, param) Failed to parse NetParameter file: train.prototxt
*** Check failure stack trace: ***
    @     0x7f8656d10daa  (unknown)
    @     0x7f8656d10ce4  (unknown)
    @     0x7f8656d106e6  (unknown)
    @     0x7f8656d13687  (unknown)
    @     0x7f8657488cd1  caffe::ReadNetParamsFromTextFileOrDie()
    @     0x7f86574e35dc  caffe::Solver<>::InitTrainNet()
    @     0x7f86574e48e3  caffe::Solver<>::Init()
    @     0x7f86574e4bff  caffe::Solver<>::Solver()
    @     0x7f86574d4eb1  caffe::Creator_SGDSolver<>()
    @           0x4108dc  caffe::SolverRegistry<>::CreateSolver()
    @           0x408da3  train()
    @           0x405c83  main
    @     0x7f865555cf45  (unknown)
    @           0x406601  (unknown)
    @              (nil)  (unknown)
Aborted (core dumped)

Can anybody training your network successfully help me? Thanks a lot!

@InnovArul
Copy link

InnovArul commented Mar 9, 2017

I think, 'VIDEO_DATA' is used in C3D-v1.0 prototxt files and it will not work in CMD-v1.1.
you may refer to https://github.com/facebook/C3D/blob/master/C3D-v1.1/examples/c3d_ucf101_training/train_simple.prototxt to rewrite your prototxt in C3D-v1.1 style.

Also, C3D-v1.1 has a Resnet inspired model. You may need to use the pretrained model given in https://github.com/facebook/C3D/blob/master/C3D-v1.1/examples/c3d_ucf101_feature_extraction/feature_extraction.sh.

@applesleam
Copy link
Author

Sorry that I didn't notice these detail. Thanks for your suggestion!

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants