AnswerDotAI / AnswerDotAI/byaldi

Models are downloaded repeatedly despite being available in cache

Open
#60 1 comment 0 reactions 0 assignees View on GitHub
Dominant language
Python
Stars
851
Forks
91
PR merge metrics
No merged PRs in 30d

Description

I am encountering an issue with the vidore/colqwen2-v1.0 model. Every time I run my script, the model downloads two versions:

models--vidore--colqwen2-v1.0
models--vidore--colqwen2-base

This occurs even though I have the two models already saved in my ~/.cache/huggingface/hub directory.
I have tried specifying the HF_HOME environment variable to point to the cache directory, but it still creates a .cache directory in the current working directory and downloads the models there.
I would like to be able to provide the absolute paths to the two model directories:
~/.cache/huggingface/hub/models--vidore--colqwen2-v1.0/snapshots/d6601e2379896aaa912c176eab8927a9b961a0e4
~/.cache/huggingface/hub/models--vidore--colqwen2-base/snapshots/fa8c8a27704510c9e4fa79b187141a1db76f50e7
However, I'm not sure how to do this correctly. Can you please help me understand how to prevent the repeated downloads and use the cached models directly?
Thank you for your assistance.

Contributor guide

No contributing guide indexed for this repository

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.