When initializing clearvoice it would be very practical to have a path parameter for the models rather than forcing current directory.
This will avoid duplicate model files in the system when running different scripts from different locations.
myClearVoice = ClearVoice(task='target_speaker_extraction', model_names=['AV_MossFormer2_TSE_16K'], models_path='/media/baredrive1/dev/clearvoice-models/')