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
Give a sidebar row one menu, and run it one step smaller (#11608)
a11beed Michael Han committed 2 months ago
Retry opening the native Safari session, and only that (#11629)
b9f8155 Daniel Han committed 2 months ago
Wait for the stored context length after reload instead of reading it once (#11624)
a468b1b Daniel Han committed 2 months ago
Pin the expression shapes a secret key may take, and read braced PowerShell env (#11621)
d6806d3 Daniel Han committed 2 months ago
Give the Docker Hub cleanup and ROCm README steps the key they read (#11619)
3a8b3e5 Daniel Han committed 2 months ago
Point the README contracts at the pages that now carry the detail, and count the drop-cue row (#11618)
1e3a9e7 Daniel Han committed 2 months ago
Keep the browser matrix TMPDIR out of the checkout (#11605)
9829e51 Daniel Han committed 2 months ago
Inject the schema fault where get_progress gets its connection (#11610)
727b033 Daniel Han committed 2 months ago
Format the two Qwen-Image-2.1 edit files the formatter check flags (#11617)
ea2a571 Daniel Han committed 2 months ago
Bump install.sh / install.ps1 pin to unsloth>=2026.9.10 (#11616)
73a8596 Daniel Han committed 2 months ago
Studio: unbreak the drag planner's diff, and key the tail where ids cannot reach (#11612)
d3909e6 Michael Han committed 2 months ago
Merge branch 'main' of https://github.com/unslothai/unsloth
219f681 Daniel Han committed 2 months ago
Studio: make sidebar drop cues say where the row actually lands (#11589)
2239158 Michael Han committed 2 months ago
Studio: show the curated model fit estimate in picker warnings (#11583)
91b525f oobabooga committed 2 months ago
Studio: install the pinned Diffusers build on startup when an old installer skipped it (#11579)
55b6b07 oobabooga committed 2 months ago
Recommend the repo's default GGUF quant instead of the largest that fits (#11606)
edd54cb Daniel Han committed 2 months ago
Probe Studio before reading the deadline in unsloth-studio-update (#11590)
7fad678 Daniel Han committed 2 months ago
Import the row fork's modules statically, which puts main back under the startup budget (#11588)
83c4c8c Daniel Han committed 2 months ago
Load Qwen-Image-2.1 GGUFs through diffusers on GPU hosts (#11604)
5c53537 Daniel Han committed 2 months ago
Do not scan a dep that only an extra installs, platform marker or not (#11582)
3b87ee8 Daniel Han committed 2 months ago
Compare a full pass to a skipping one without the click both record as unmet (#11578)
20f3f50 Daniel Han committed 2 months ago
Expect Dense (bf16) after a declined encoder precision (#11580)
d5dc3ff Daniel Han committed 2 months ago
Baseline the seven unsloth-zoo 2026.9.7 findings after re-review (#11577)
321a4ef Daniel Han committed 2 months ago
Keep scrolling to the end of the reasoning transcript in the thread settings driver (#11575)
e8d6ef9 Daniel Han committed 2 months ago
Let the composite renaming tests run inside the transformers 5.4 window (#11572)
30868af Daniel Han committed 2 months ago