search

facebookresearch / Noresqa

This github repo is for Neurips 2021 and Interspeech 2022 papers on Non-Matching Reference based estimation of speech quality assessment.
Other
88 stars 12 forks source link

The pre-trained model of "model_noresqa_mos.pth" can not be successfully loaded. #8

Closed xiaoyiming closed 2 years ago

xiaoyiming commented 2 years ago

model_checkpoint_path = 'models/model_noresqa_mos.pth' state = torch.load(model_checkpoint_path,map_location="cpu") Traceback (most recent call last): File "", line 1, in File "C:\ProgramData\Anaconda3\envs\xymtest2\lib\site-packages\torch\serialization.py", line 593, in load return _legacy_load(opened_file, map_location, pickle_module, pickle_load_args) File "C:\ProgramData\Anaconda3\envs\xymtest2\lib\site-packages\torch\serialization.py", line 763, in _legacy_load magic_number = pickle_module.load(f, pickle_load_args) _pickle.UnpicklingError: invalid load key, 'v'.

argxkr commented 2 years ago

I am not able to reproduce this error. What us the pytorch version you are using ? We have tested with 1.12.0 and loading works.

xiaoyiming commented 2 years ago

I am not able to reproduce this error. What us the pytorch version you are using ? We have tested with 1.12.0 and loading works.

Thanks, my torch version is 1.7.0. I will try with 1.12.0

xiaoyiming commented 2 years ago

I am not able to reproduce this error. What us the pytorch version you are using ? We have tested with 1.12.0 and loading works.

It seems that there is a problem with the uploaded pre-training file, please check whether the complete file has been uploaded image

dv-fenix commented 2 years ago

@xiaoyiming I was facing the same issue earlier. The model files have been uploaded using Git Large File Storage (git-lfs). I would recommend that you either download them separately or use git lfs to clone the repository.

argxkr commented 2 years ago

I think cloning with git lfs should work. I am closing this issue for now.