Samsul Rahmadani (munggok)

yeah me too longer training,,the voice is clearer and similar, but cant pronounce a single word maybe phenomizer problem ,idk

@jeremy110 yes all my tone are set to 0 now wondering how can i fix this

> The script that does the filtering can be found here: https://github.com/neonbjb/DL-Art-School/blob/master/codes/scripts/audio/preparation/pipeline.py > > You will need a few files: > > 1. Classifier weights, [here](https://drive.google.com/file/d/1Mg5ok2Es7utGDvWzwX_pjugDKzhrTT9q/view?usp=share_link) > 2. Classifier config,...

@neonbjb thanks for the reply i run the code with cvvp config but found error (`KeyError: 'clip'`) but i found your another repo : https://huggingface.co/jbetker/tortoise-filtering-models/tree/main, use yaml and model config...

also encounter this with python 3.11 > CFLAGS="-msse3 -msse4.1 -msse4.2 -mavx -mavx2" pip install nmslib==2.1.1 tried this, not work colab to recreate: https://colab.research.google.com/drive/1R-iVww3-4zAuzWrqecsJIG0x1tEUaqLK?usp=sharing

hello @neonbjb want to ask about data format especially about podcast when you split the the audio( i assume),,do you make sure for each split contain only 1 speaker,or is...

+1 to this i'm already fine tuning bloom for instruction task and the result quite good bloomz.cpp seem doesnt have capabilities to do inference for instruction task (define prompt)

@matthayes i'm just changing this line to model that i want https://github.com/databrickslabs/dolly/blob/03bf3852daa42e6091a39483dda0714c02de7382/training/trainer.py#L35 so far i tried [bloom](https://huggingface.co/bigscience/bloomz-7b1-mt) and xglm,, got same error

+ 1

@gwenzek this also happen in dumps 2020-24 , 2020-29, 2020-40, 2020-45, 2020-50,