94bb494nd41f

Results 4 comments of 94bb494nd41f

LoRA Training CPU not working, llama_tokenize: too many tokens

This is my first issue, some how mine looks different?

ImportError: cannot import name 'download_loader' from 'llama_index'

switching to python 3.9.16 avoids the error

Crash on low end PC

What kind of model were you trying to load? I guess 8 GB is not that much but should work fine with smaller modells like https://huggingface.co/TheBloke/wizardLM-7B-GGML although even the smalles...

Query giving wrong citations and answers out of the document

I guess the Modell is just reproducing "knowledge" it has learned. Since the results do not fit the question but are the best available, it might still answer. How to...