Drew Thomasson

The temp fix workaround I found was instead of using print(TTS().list_models()) I do: from TTS.api import TTS tts_manager = TTS().list_models() all_models = tts_manager.list_models() print(all_models) doing that makes it actually print...

I thought it was if the model has multi-dataset in the name of the model when you make it print out all of the tts models? #Python print(TTS().list_models()) Or I...

hm, In that case I'm all for the feature your requesting

I also await this fix!

Tried that, been doing more digging, installing cuda all that jazz, it seems that ill just need a gpu to run unsloth :/ Well there's the google colab anyway so...

does `pip install tensorflow-metal` also work? hoping that can give it the neural chip speed up of the apple silicone chip

I updated the instruction cause the transformers version doesn't work. `pip install tensorflow-macos` `pip install torch` `pip install spacy` `pip install transformers==4.30.0` `pip install --no-dependencies booknlp` and it should run...

case no one else I made a booknlp pip install specifically for apple silicone! No npu or gpu speedup tho :( `pip install apple-silicon-booknlp` just make sure its in a...

Seems like it runs well in WSL for windows, it even can use cuda. You can access the files on your wsl ubuntu in windows file exploror by putting this...

considering this never fully compiled for window I made a pip install for windows! seems to be working on my end `pip install windows-booknlp`