Local UI to run and train LLMs and diffusion models. Supports GGUF, MLX, Qwen3.8, DeepSeek-V4.1, MiniMax-H3, Gemma 4, FLUX and more.
Python agentaichatgptdeepseekfine-tuninggemmaimage-generationllamallmllmsopenaipythonqwenreinforcement-learningself-hostedstable-diffusiontext-to-speechttsuiunsloth
Commits on main
Stop dividing the GRPO eval loss by the accumulation steps (#11337)
06c1bb4 Vineeth Sai Varikuntla committed 2 months ago
Keep credentials out of cached directories, and audit the org against the miri cache disclosure
43e7dd3 Daniel Han committed 2 months ago
Studio: read the scaled lengths and colours the UI contracts were written against (#11505)
655a195 Daniel Han committed 2 months ago
Fix the update card clipping its own buttons at the 20px interface size (#11510)
de71b32 Daniel Han committed 2 months ago
Cut CI queue: scope push triggers, and rebalance Repo tests (CPU) on measured runner time
a78110b Daniel Han committed 2 months ago
Unsloth Studio: make a newly created API key selectable when the clipboard is unavailable (#11489)
a9f5348 Leo Borcherding committed 2 months ago
Stop unsloth chat and unsloth inference reloading a 16-bit model in 4-bit (#11477)
3c3475e Nilay committed 2 months ago
Studio: show what the caches cost and let them be cleared (#10528)
f183736 Daniel Han committed 2 months ago
Stop the installers importing the download stack just to be imported (#11446)
edf77d7 Daniel Han committed 2 months ago
Repair a quarantined llama.cpp runtime at launch instead of failing at model load (#10494)
acf2de6 Daniel Han committed 2 months ago
studio/frontend: keep shutdown dialog open when server-stop fails (#5637)
cd8f87f Daniel Han committed 2 months ago
Studio: promote -ts into tensor_split before manual strip (#11334)
8a9ad3c Indrajit swain committed 2 months ago
fix(studio): track the repository resolved by the training loader (#11462)
b490d56 Lee Jackson committed 2 months ago
Studio: keep a streaming code fence highlighted, and window its spans (#11423)
40ade0f Daniel Han committed 2 months ago
Studio: keep <think> text visible in the reply when Thinking is off (#11481)
9055780 Nilay committed 2 months ago
Stop unsloth chat and unsloth inference ignoring the model's recommended settings (#11478)
5708db9 Nilay committed 2 months ago
Make the drag preferences test compile, and ask the stricter question (#11503)
5e04590 Daniel Han committed 2 months ago
Studio: make the contrast slider reach surfaces and controls (#11459)
3bc5eb5 Michael Han committed 2 months ago
Anchor the worker source read on the test file, not the working directory (#11495)
42fc28f Daniel Han committed 2 months ago
fix(studio): bundle hf-xet for Desktop large Hub downloads (#10849)
c052e1f Sourav Rajvi committed 2 months ago
Studio: spell the owner's sidebar name Unsloth, not unsloth (#11275)
b5cb7bb Michael Han committed 2 months ago
Studio installers: add --no-rollback, and report a full disk as a full disk (#11364)
62923f2 Daniel Han committed 2 months ago
Studio: give back the continuation lease when the user stops the preflight (#9560)
9ca2930 Daniel Han committed 2 months ago
Read the 16-bit pixels the way the rest of this file already does (#11493)
1c9db26 Daniel Han committed 2 months ago
Give the GPU summary probe test the helpers its extracted block calls (#11472)
726c32a Daniel Han committed 2 months ago
Pin the worker RSS in the two xet ledger tests that stand in for a worker (#11492)
826549e Daniel Han committed 2 months ago
studio: copy full prompts from the api monitor (#11340)
a9feb95 Maheswar Kumar committed 2 months ago
Pin the hardware verdict in the two export capability tests that depend on it (#11490)
259a506 Daniel Han committed 2 months ago
Keep the confinement tests off the interpreter walk (#11488)
461f5ab Daniel Han committed 2 months ago