New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Unable to import mobilenet model using latest OpenCV. #9462
Comments
First thing I saw that you have to use fused batch normalization. Otherwise graph has constant operations. TensorFlow has multiple different implementations for some of layers. Unfortunately, only part of them might be parsed easily. In example, only For now, you may find a test case with supported BN here: https://github.com/opencv/opencv_extra/blob/master/testdata/dnn/tensorflow/generate_tf_models.py#L85, but I think we can prepare some tool to transform the graphs into readable format. |
Thank you very much, I will try the test case now. It will be wonderful if there are some tool to transform the graphs of mobilenet, thanks again! |
|
I've landed here because I got the same error while trying out a mobilenet 128 graph. It loads fine by the 'label_image' example, but fails with the OpenCV importer, giving the same error JackGan999 reported. |
@JackGan999, @ioannisnousias, may I ask you to test the changes from PR #9517? However I used a public checkpoint. |
System information (version)
output_labels.txt
mobilenet_1.0_224.pb.tar.gz
The text was updated successfully, but these errors were encountered: