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 user report filed against LM Studio says two gaps in version 1.1.8 make some decision models difficult to use through its API. The report is an unconfirmed user account, not a confirmed product defect.

Why it matters

The user says a model marked as decision-capable fails to load in the bundled runtimes, while the load API cannot set the physical batch size needed for longer prompts. In the example provided, a 2,845-token prompt is rejected at the default batch size of 512, even when a higher evaluation batch size is requested. For developers building around LM Studio’s decision endpoints, the distinction between those settings could be the difference between a model that appears configurable and one that handles a real workload. The report is a useful lead, pending a response or confirmation from LM Studio.

Discuss: Should local-model tools expose every inference setting through their APIs, or is it reasonable to leave some configuration in the desktop app?

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.