lmstudio-ai / lmstudio-ai/mlx-engine
MLX Runtime Visual Model Multiple Image bug.
@neilmehta24 is already working on this.
Since May 5, 2025.
- Dominant language
- Python
- Stars
- 1.2k
- Forks
- 133
- Avg merge
- 21h 6m
- Merged PRs (30d)
- 1
Description
When using visual models with the MLX backend, an error occurs after sending more than two images, causing the model to incorrectly recognize images. The model used is qwen2.5-vl-32b-instruct-6bit, and this issue also appears on other visual models like mistral-small3.1. The hardware being used is M3 Ultra, and this problem exists in MLX versions 0.14.0 and earlier versions such as 0.13.2.
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Assessment
This issue has not been assessed yet.