feat: configurable model name, description and vision capability

[cheahjs]

Pull Request Checklist

• Target branch: Pull requests should target the dev branch.
• Description: Briefly describe the changes in this pull request.
• Changelog: Ensure a changelog entry following the format of Keep a Changelog is added at the bottom of the PR description.
• Documentation: Have you updated relevant documentation Open WebUI Docs, or other documentation sources?
• Dependencies: Are there any new dependencies? Have you updated the dependency versions in the documentation?
• Testing: Have you written and run sufficient tests for the changes?
• Code Review: Have you self-reviewed your code and addressed any coding standard issues?

---

Description

Adds the ability to set a custom display name, description, and the vision capability of a model via the web UI, which is then persisted into the database.

A new field custom_info is injected into each model list API. An alternative implementation would be to override the model's name field, but would require additional mapping inside of the proxy endpoints.

If a model is marked as not vision capable,

1. Image uploads are disabled.
2. Any existing images that are in the current input are removed - but not any images that are already in the chat
3. If multiple models are selected, and not all of them are vision capable:
   1. A warning icon is added to image input warning about this
   2. An error toast is shown if a user still sends the prompt.

I've tested locally with Ollama and OpenAI endpoints, have not tested LiteLLM endpoints yet.

Changelog

Added

• Model Customization: Configure custom display names, descriptions, and vision capabilities for models.
• Image Input Validation: Models marked as not vision-capable will reject or warn about image uploads.

[justinh-rahb]

LEGENDARY.

[tjbck]

About time we make config.json persistent as well, I'll investigate this.

[tjbck]

Also, it would be perfect if we could rename the fields like such:

{
  "id": "llama3:latest",
  "name": "Meta: Llama3 8B (Instruct, Q3, K, S)",
  "description": "Llama3 is a large language model trained on a diverse range of internet text.\nIt is capable of generating human-like text and answering a wide range of questions.",
  "vision_capable": false
},

[cheahjs]

Done the field rename, and also added settings UI for updating the model info.

[Yanyutin753]

It is so unbelievable. I've been waiting for a long time! @cheahjs i admire you so much!!!🤩🤩🤩

[arsaboo]

This is epic. I am assuming we can extend this (in a separate PR) to configure per-model context window, output tokens, etc.

[tjbck]

On second thought, we might want to save the model configs to our database to cover #665 as well. I'll keep this open in case I change my mind. I appreciate the effort as always 🙌