search

mudler / LocalAI

:robot: The free, Open Source alternative to OpenAI, Claude and others. Self-hosted and local-first. Drop-in replacement for OpenAI, running on consumer-grade hardware. No GPU required. Runs gguf, transformers, diffusers and many more models architectures. Features: Generate Text, Audio, Video, Images, Voice Cloning, Distributed inference
https://localai.io
MIT License
24.32k stars 1.86k forks source link

Identify model by type (tts, text, ...) #3488

Open ecyht2 opened 1 month ago

ecyht2 commented 1 month ago

LocalAI version: I am using the docker image from dockerhub localai/localai:lastest-cp but I don't think it matters.

Environment, CPU architecture, OS, and Version: Linux archserver 6.6.49-1-lts #1 SMP PREEMPT_DYNAMIC Wed, 04 Sep 2024 13:18:51 +0000 x86_64 GNU/Linux

Describe the bug Models are appearing where they shouldn't be in "Select Model". For example, a text model is showing up in TTS, Image Generation, and Talk. image

To Reproduce

  1. Install a text model.
  2. Go to TTS tab.
  3. Click "Select Model" combo box.
  4. Text model available there.

Expected behavior Text models should only appear in the "Chat" tab and vice versa for other models.

Logs N/A

Additional context N/A

dave-gray101 commented 1 month ago

I may not be remembering correctly, but I don't believe we currently filter these lists. I have a PR pending review that will add a heuristic function able to guess which use cases are appropriate for each dropdown.

Please note that this pr doesn't actually extend support to changing these dropdowns yet, as I'd like to run a few tests and experiments once the unit test is merged first.

I'll punch in a quick cli command to exercise the functionality in case others are able to help test the functionality against different model configurations to speed this along

mudler commented 1 month ago

we don't filter by model category yet - but this is something in our radar and that should be addressed at some point