Prince Canuma comments

Results 572 comments of


                                            Prince Canuma

Keeping track of the performance and compatibility of models

Could you be a bit more specific? Model name + reproducible example.

Keeping track of the performance and compatibility of models

Also past success on the same example.

Keeping track of the performance and compatibility of models

Thanks a lot @jrp2014!

KeyError: 'image_token_index'

> Maybe Qwen/Qwen2.5-VL-7B-Instruct requires additional support? Hey @pavelgur Thanks for reporting this. It's a small bug in the MoldelConfig because Qwen2.5 uses `image_token_id` instead of the commonly used `image_token_index`. I...

KeyError: 'image_token_index'

Will fix in a upcoming release

Failed to use fastvlm to inference due to arguments not match

Hey @hermeschen1116 Thanks for reporting it! I'm gonna do a refactor because this is caused by mlx-lm changes

Failed to use fastvlm to inference due to arguments not match

could you share the issue?

Failed to use fastvlm to inference due to arguments not match

Could you share a reproducible example?

Support Chat API

You can do this manually. I will later check if there is a great way to integrate this. The problem is that not all models support this format.

Support Chat API

Also we have a multi-turn example here: https://github.com/Blaizzy/mlx-vlm/issues/68#issuecomment-2440233679