Colab Notebook for easy use

Hello,

I've created a notebook for Google Colab to use your great project in a more convenient way:
https://github.com/jarredou/Apollo-Colab-Inference

I've created a new inference.py file with chunking and overlapping features to allow seamless processing of long audio files.

[JusperLee]

Thank you so much for your help. I'm bumping your issue for other users.

[bejucoo]

Hello, I was trying to run the notebook and got the following error:

In PyTorch 2.6, we changed the default value of the weights_only argument in torch.load from False to True. Re-running torch.load with weights_only set to False will likely succeed, but it can result in arbitrary code execution. Do it only if you got the file from a trusted source.

So in the Apollo/look2hear/models/base_model.py file I changed line 64 to:

conf = torch.load(pretrained_model_conf_or_path, map_location="cpu" , weights_only=False)

Seems like is running fine now. Another way could be installing "torch<2.6.0".

@bejucoo force torch < 2.6 version install to fix that issue for now.

I'll edit the colab notebook in the coming days.