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
Studio: don't take a port another process holds on 0.0.0.0 (#12011)
da58421 Suliman Abdulrazzaq committed 1 month ago
Studio: keep CUDA 12 llama.cpp prebuilts off drivers older than CUDA 12.4 (#13031)
0461328 Daniel Han committed 1 month ago
Studio: restart llama-server after a lost GPU device instead of locking the chat (#13044)
fc56663 Daniel Han committed 1 month ago
Studio: install the released Diffusers 0.41.0 instead of the pinned main commit (#13045)
7730160 Daniel Han committed 1 month ago
Studio: re-measure annotate marks after a tab zoom commits (#13082)
3f07e7f Michael Han committed 1 month ago
Save models loaded from pre-quantized bitsandbytes checkpoints on transformers 5 (#13054)
cfdb966 Daniel Han committed 1 month ago
Studio: int8 ConvRot Qwen-Image-2.1 text encoder on Apple Silicon, and release text encoders during denoise on unified memory (#13026)
97b416c Daniel Han committed 1 month ago
Keep loss-only evaluation on the fused CE path instead of materializing logits (#13065)
b799e4d Daniel Han committed 1 month ago
Bump install.sh / install.ps1 pins to unsloth>=2026.10.3 and unsloth-zoo>=2026.10.3 (#13083)
f0a2c7b Daniel Han committed 1 month ago
Studio: point compressed-tensors, AWQ and GPTQ models to the vLLM engine (#13043)
d94ca0e Daniel Han committed 1 month ago
Train EXAONE 3.5: name the token embedding remote code no longer exposes (#13058)
f32da6a Daniel Han committed 1 month ago
Merge branch 'main' of https://github.com/unslothai/unsloth
69524f8 Daniel Han committed 1 month ago
Studio: load a repo id from its scan-folder copy instead of re-downloading it (#13034)
bcdc3d2 Daniel Han committed 1 month ago
Merge branch 'main' of https://github.com/unslothai/unsloth
c26e84b Daniel Han committed 1 month ago
Studio: keep a dragged annotate area the size it was drawn (#13070)
b0f0107 Michael Han committed 1 month ago
Studio: name media companion downloads by component, and show what Run still fetches for an on-device GGUF (#13027)
345ff9a Daniel Han committed 1 month ago
Studio: report the real llama.cpp version for source builds (#13036)
56a6ecd Daniel Han committed 1 month ago
Studio: pin FastFlowLM 1.0.7 for Qwen3.8 27B on the AMD NPU and keep Lemonade / FastFlowLM current (#13047)
d7bfce6 Daniel Han committed 1 month ago
Studio: serve decision models through the MLX engine on Apple Silicon (#13015)
fde93c4 Long Yixing committed 1 month ago
Studio: Docs links for Sandbox, Agents and the Decision API (#13076)
132dafa Michael Han committed 1 month ago
Make Q-GaLore optimizer state resumable from checkpoints (#13051)
6be1262 Daniel Han committed 1 month ago
Allow backward through eval-mode and for_inference forwards (#13050)
495aedd Daniel Han committed 1 month ago
Studio: keep tensor split and MTP when Auto would pick a DFlash drafter that aborts it (#13040)
bc915c0 Daniel Han committed 1 month ago
Installer: read the venv's torch in isolation so a PYTHONPATH torch cannot break the install (#13041)
6d3af3c Daniel Han committed 1 month ago
Studio: give the annotate comment box a visible shadow in dark mode (#13071)
9f447a4 Michael Han committed 1 month ago
Studio: let another account join the resident GGUF instead of replacing it (#13038)
2e80b73 Daniel Han committed 1 month ago
Decline the packed INT4 kernels when weight_scale has the wrong group count (#12957)
0cb4066 Jay Zhou committed 1 month ago