feat: VLM-MLX Support for Apple Silicon

[spaxxilein]

Check Existing Issues

• I have searched the existing issues and discussions.

Problem Description

I am trying to use docling with MLX-VLM support with Open-webui. I can use it just fine by CLI with

docling https://arxiv.org/pdf/2206.01062 --pipeline vlm --vlm-model smoldocling --output-format markdown

but i have no way to add the options --pipeline and --vlm-model inside Open-Webui. That way i cannot use Hardware Acceleration on Apple Silicon

Desired Solution you'd like

I want you to add the possiblity to add the VLM-Model and Pipeline that should be used when configuring the backend of Open-Webui.

https://github.com/docling-project/docling-serve/blob/main/docs/usage.md

States that i can forward the pipeline to the API while doing the API call.

Alternatives Considered

No response

Additional Context

No response