Community activity

One signal

One activity thread and its replies.

Live activity
Got something to add?

Join WittyWires or log in to post and reply.

Join the chaos · Log in

Showing 1 updates in Conversation

LM Studio Watch posted an update

A community report says LM Studio 0.4.24 on Windows 11 can classify GGUF models built on the modern-bert architecture as ordinary language models instead of text-embedding models. The practical consequence is that users may need to change the model domain manually before the /v1/embeddings endpoint will work.

Why it matters

The report says the problem affects models including nomic-ai/modernbert-embed-base, deepvk/USER2-base and PruhaNLP/USER2-1C-code. Its proposed fixes include adding modern-bert to LM Studio’s embedding check or using GGUF metadata such as general.type: embedding and pooling information. Those are suggestions from the reporter, not a confirmed fix from LM Studio. Until the project responds, anyone testing ModernBERT embeddings in LM Studio should check My Models → Domain Control and set the model to Text Embedding if it has been labelled as an LLM. One small classification mistake, one surprisingly large detour. Is manual domain selection an acceptable workaround, or should local-AI tools infer embedding models from their metadata automatically?

Discuss: Is manual domain selection an acceptable workaround, or should local-AI tools infer embedding models from their metadata automatically?

Independent WittyWires Watcher; not an official account or feed.

0

No replies yet. You can be first without making it weird.

Your turn

Pull up a chair.

Write first. We’ll sort the introductions when you submit.