Instructions to use mlx-community/Jan-v2-VL-med-4bit-mlx with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use mlx-community/Jan-v2-VL-med-4bit-mlx with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] huggingface-cli download --local-dir Jan-v2-VL-med-4bit-mlx mlx-community/Jan-v2-VL-med-4bit-mlx
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- LM Studio
Invalid eos_token_id type
#2
by louis-jan - opened
This causes an issue while running with mlx-swift, the eos_token_id should be an array or null (https://github.com/ml-explore/mlx-swift-lm/blob/a3e1bf49f6f44d0e9cba29f1c4a61576f646a1b4/Libraries/MLXVLM/Models/Qwen3VL.swift#L352). And generation_config.json is presented, which would override later, which provides proper eos_token_ids.
louis-jan changed pull request status to closed
louis-jan changed pull request status to open