unsloth/studio/backend/core
Roland Tannous 43a633e550 feat: boot llama-server with --parallel 4 by default
Raise the default parallel slot count so Studio chat and the external API
endpoint can run concurrently without a subprocess restart. Enable/disable
of the Access Endpoint is now instant — it only mints or clears the API
key, since the parallel slots are already in place from the initial load.
2026-04-09 14:56:01 +00:00
..
data_recipe build(deps): bump oxc-parser (#4776) 2026-04-08 03:35:33 -07:00
export split venv_t5 into tiered 5.3.0/5.5.0 and fix trust_remote_code (#4878) 2026-04-07 20:05:01 +04:00
inference feat: boot llama-server with --parallel 4 by default 2026-04-09 14:56:01 +00:00
training split venv_t5 into tiered 5.3.0/5.5.0 and fix trust_remote_code (#4878) 2026-04-07 20:05:01 +04:00
__init__.py Combine studio setup fixes: frontend caching, venv isolation, Windows CPU support (#4413) 2026-03-18 03:52:25 -07:00