Pretraining Hubert

[shirounanashi]

What is your question?

I'm trying to train a Hubert model, and I'm having the problem that the .km files are coming out in ANSI instead of UTF-8. I'm using colab to create these files

Code

https://colab.research.google.com/drive/1Ju6sLIoHhIl5LEtyaWMuDuryDt7qbGJs?usp=sharing

What have you tried?

Training a hubert model

What's your environment?

• fairseq Version (e.g., 1.0 or main): 1.22.2
• PyTorch Version (e.g., 1.0): Default in colab
• OS (e.g., Linux): Linux
• How you installed fairseq (pip, source): pip
• Build command you used (if compiling from source):
• Python version: 3.10
• CUDA/cuDNN version:
• GPU models and configuration: T4
• Any other relevant information:

[shirounanashi]

Forget it, I didn't run part of the code that was necessary to run