MIT

/

ast-finetuned-audioset-10-10-0.4593

like 147

Hi,

There seems to be something wrong with your environment, would it be possible to create a fresh virtual environment, install PyTorch and Transformers there and see if you're still facing the error?

Failed to import transformers.models.audio_spectrogram_transformer.feature_extraction_audio_spectrogram_transformer because of the following error (look up to see its traceback): libc10_cuda.so: cannot open shared object file: No such file or directory

It seems like an error with your CUDA installation. you can check it here:

https://xcat-docs.readthedocs.io/en/stable/advanced/gpu/nvidia/verify_cuda_install.html

Then, you can do a run a simple test in python to assess that torch has access to the GPU (this should print a 10x10 tensor filled with ones that are allocated in the cuda device):
python -c "import torch; print(torch.ones((10,10)).to('cuda'))"

Hope it helps.