Skip to content

fixing exports in preprare llm models script#4340

Open
przepeck wants to merge 2 commits into
mainfrom
przepeck/fix_exports_preprare_llm_scripts
Open

fixing exports in preprare llm models script#4340
przepeck wants to merge 2 commits into
mainfrom
przepeck/fix_exports_preprare_llm_scripts

Conversation

@przepeck

Copy link
Copy Markdown
Collaborator

🛠 Summary

CVS-189149
Fixing prepare_llm_models.sh by changing order of exports and changing version of typer python lib.

🧪 Checklist

  • Unit tests added.
  • The documentation updated.
  • Change follows security best practices.
    ``

Copilot AI left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

This PR updates prepare_llm_models.sh, which provisions a local model repository for OVMS LLM-related tests/demos, by adjusting how some tokenizers are obtained and by reordering the VLM preparation block.

Changes:

  • Added a new model constant (LFM25_MODEL).
  • Switched GEMMA4 tokenizer preparation from hf download --include *tokenizer* to convert_tokenizer --with_detokenizer.
  • Moved the VLM block later in the script and introduced an on-demand typer upgrade before running hf download.

Comment thread prepare_llm_models.sh
Comment thread prepare_llm_models.sh
Comment on lines +226 to +228
pip3 install --upgrade typer==0.25.1
hf download "$VLM_MODEL" --local-dir $1/$VLM_MODEL
convert_tokenizer OpenGVLab/InternVL2-1B --with_detokenizer -o $1/$VLM_MODEL # WA to use newer tokenizer model format which supports padding.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants