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: keep loop notes ahead of the MCP image envelope (#11367)
60c86b9 Jyotirmoy Roy committed 1 month ago
Studio: gate auth.db and rag.db opens and closes against the SQLite 3.51.0-3.51.1 deadlock (#13307)
40eb194 Daniel Han committed 1 month ago
Studio: pair torch <= 2.8 with torchao 0.15 and gate torchao export on 0.15 (#13329)
3a1b694 Daniel Han committed 1 month ago
torchao alias tests: stop leaving a fake diffusers in sys.modules (#13336)
41485d3 Daniel Han committed 1 month ago
Studio: remember Max Tokens on Max as Max, not as the old context length (#13223)
2afec3c Daniel Han committed 1 month ago
Apply an explicit rope_scaling to the model and to vLLM (#13315)
7d2bd01 Daniel Han committed 1 month ago
Intel XPU: 4bit loads without a C compiler by falling back to bitsandbytes' PyTorch quantize (#13309)
956430a Daniel Han committed 1 month ago
Studio: keep llama.cpp's Metal fit inside the GPU wired headroom (#13319)
548fe07 Daniel Han committed 1 month ago
Studio: reload without MTP when llama-server dies mid-generation on any MTP launch (#13316)
b5efe36 Daniel Han committed 1 month ago
Studio: fix torch.compile for image generation on AMD (training config leak, Windows without Visual Studio, max tier) (#13227)
0acf8b1 Daniel Han committed 1 month ago
Studio: price the int8 ConvRot text encoder on Apple Silicon when the dense one is still cached (#13306)
c860f29 Daniel Han committed 1 month ago
Unsloth Studio: export a trained image LoRA from the Train tab, list exported copies once (#13289)
e5bae1a Daniel Han committed 1 month ago
Studio (Windows): read a runtime lock held by an elevated Unsloth process as busy, not a broken install (#13301)
44d6349 Daniel Han committed 1 month ago
Studio: torch 2.13 for new Windows x64 cu130 Python 3.13 installs, existing installs keep their torch (#13263)
9c4c4a1 Daniel Han committed 1 month ago
Studio: install prebuilt flash-attn, causal-conv1d and mamba-ssm wheels on Windows and ARM (#13262)
14ad75e Daniel Han committed 1 month ago
Build the prebuilt CUDA 13 kernel wheels for Windows and ARM (win_amd64, linux_aarch64, win_arm64) (#13261)
0fbda96 Daniel Han committed 1 month ago
Studio installer (macOS): start the Unsloth Studio.app shortcut natively on Apple Silicon (#13305)
2640084 Daniel Han committed 1 month ago
Studio: fix image and video loads on torchao 0.14 (FqnToConfig import error with diffusers 0.41) (#13308)
ec65aaf Daniel Han committed 1 month ago
Run GRPO under FSDP2: no inference mode around FSDP, an optimizer that steps DTensors, and refuse vLLM with FSDP2 (#13156)
cc10f80 Daniel Han committed 1 month ago
Studio: draw browser menus, tooltips and toasts over a live native page (#13122)
f7f81ff Michael Han committed 1 month ago
Studio: shared media bar, Open with apps, steady image lightbox (#13241)
ba21f79 Michael Han committed 1 month ago
Studio: pair Qwen3.5 / 3.6 vision projectors from another quantizer (#13288)
93f1dbd Daniel Han committed 1 month ago
Studio: stop dropdowns jumping on open under Reduce Motion (#13238)
323e0a8 Michael Han committed 1 month ago
Studio: move chat with files into a panel above the composer (#13078)
00958e6 Michael Han committed 1 month ago
Studio: bound the ChatGPT subscription and NPU disconnect watcher teardown (#13293)
82850ba Daniel Han committed 1 month ago
Studio: install and run vLLM on AMD GPUs, on Linux and through WSL on Windows (#12785)
c54ab45 oobabooga committed 1 month ago
Studio: name an MLX or managed-engine token count by the id status publishes (#13282)
fc5cf50 Daniel Han committed 1 month ago
Run packed rows on the transformers modeling path through varlen flash attention (#13234)
c554fbf Daniel Han committed 1 month ago