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
Let the Projects section stand in for its row in the macOS tab walk (#12074)
1cecdb4 Daniel Han committed 1 month ago
Reset the permission step's storage at the start of the next document (#12073)
7a22515 Daniel Han committed 1 month ago
Studio: keep the sidebar menu shadow in light mode (#12062)
c807acb Michael Han committed 1 month ago
Turn off Dr GRPO reward scaling under TRL's "group" default (#11951)
c0b27d2 Vineeth Sai Varikuntla committed 1 month ago
Count #12016's section header among the sidebar's scaled 30px rows (#12066)
a37b610 Daniel Han committed 1 month ago
Give the health wait's working-child tests room for the worker to start (#12065)
5d1c588 Daniel Han committed 1 month ago
Skip chordless menu items in the native chord collision check (#12063)
c595f22 Daniel Han committed 1 month ago
Studio: custom sidebar sections, section menus and drag to reorder (#12016)
b47fb09 Michael Han committed 1 month ago
Studio: let Deep Research finish a turn handed off from a chat generation (#11923)
d7eba2e Mohammad Hijjawi committed 1 month ago
Let the macOS tab sampler ride out a navigation still in flight after login (#12053)
d294f79 Daniel Han committed 1 month ago
Run the formatter fixed-point guard's batches side by side (#12050)
a16cac1 Daniel Han committed 1 month ago
Keep peft's is_torchao_available cache API through the stale-torchao patch (#12049)
2465ef2 Daniel Han committed 1 month ago
Advertise NVFP4 diffusion on the text encoder driver's mocked host (#12047)
8faa867 Daniel Han committed 1 month ago
Resolve the macOS app menu's chords as a Mac in its test (#12046)
e05cdd9 Daniel Han committed 1 month ago
Read the partial safetensors delete-menu guard by operator, not verbatim (#12042)
ab41b1c Daniel Han committed 1 month ago
Translate the Library toolbar Sort menu in every locale (#12045)
0ddd03e Daniel Han committed 1 month ago
Unsloth Studio: report the VAE decode on the image progress bar (#11740)
195d8dd Leo Borcherding committed 1 month ago
Studio: stop trading the denoiser's quantisation away to pay for offload (#11558)
b6ee173 Daniel Han committed 1 month ago
Keep flash attention off sub-models that do not support it (LFM2.5-VL SigLIP2 tower) (#11959)
6a4af10 Daniel Han committed 1 month ago
Studio: fall back to native when SageAttention does not run on this GPU (#11997)
f5d3cd3 Daniel Han committed 1 month ago
Grouped-linear LoRA for DeepSeek-V4, remote-code shims for Step-3.7, and a real message for Mistral-format checkpoints (#11528)
893889d Daniel Han committed 1 month ago
Studio: opt-in Hadamard rotation for Qwen-Image-2.1's int8 transformer so it matches bf16 (#11835)
b33abbc Daniel Han committed 1 month ago
SentenceTransformer: add opt-in FP32 merged-pair ranking loss (#11867)
676e09f Etherl committed 1 month ago
Studio: pass tool_result is_error through to the model on /v1/messages (#11962)
324f4c0 Nilay committed 1 month ago
Put back the unsloth_zoo modules the device map opt-in tests stub (#12038)
8c2ffee Daniel Han committed 1 month ago
Load only the language model for text_only on repo-code composites (#11861)
751dbe7 Daniel Han committed 1 month ago
Apply SFTConfig.router_aux_loss_coef to MoE models that cache it at init (#12006)
c3c388b Daniel Han committed 1 month ago
Unsloth Studio: stop offering "Continue" on a base repo that only holds a GGUF's text encoder and VAE (#11644)
5586618 Leo Borcherding committed 1 month ago