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 gpt-oss generation at the harmony tool call token (#11449)
7bacc24 Daniel Han committed 1 month ago
Studio: name the MCP server while the tool call is still streaming (#9212)
329a7de Daniel Han committed 1 month ago
Load Mistral-format checkpoints (params.json only) through transformers' Mistral4 (Mistral-Large-3) (#12144)
78afe44 Daniel Han committed 1 month ago
Do not reinstall llm-compressor when it is already installed (#6806)
949b94a Daniel Han committed 1 month ago
Studio: return no-op instead of 500 for an out-of-range scan folder id (#8397)
94dda36 Daniel Han committed 1 month ago
Unsloth Studio: stop a Mac browser hiding GPU-only models from a remote CUDA / ROCm / Intel server (#8833)
60bf93c Daniel Han committed 1 month ago
Studio: compile VAEs by repeated block, one H3 graph per first render, vectorise the HunyuanVideo-1.5 VAE mask (#12061)
0771bdc Daniel Han committed 1 month ago
Studio: compile FLUX.1 dynamic, CUDA-graph the SDXL U-Net, decode one-frame Qwen-Image latents in 2D (#12060)
c6eb2e6 Daniel Han committed 1 month ago
Load block-FP8 checkpoints in 4-bit (NF4) when load_in_4bit=True is passed (#12146)
91bcb18 Daniel Han committed 1 month ago
Studio: fix inductor CantSplit on torch 2.12/2.13 and speed up the Qwen-Image-2.1 VAE (#12059)
03d3cc4 Daniel Han committed 1 month ago
Studio: capability tags in the model picker match the vision pill and get their own colours (#12192)
da85782 Daniel Han committed 1 month ago
Studio: fused int8 MLP kernels and real-arithmetic RoPE for DiT denoisers (FLUX.1 11% less GPU time per step) (#12083)
8f615ac Daniel Han committed 1 month ago
Count only the retry loop's own sleeps in the JSON fallback backoff test (#12189)
9970dcc Daniel Han committed 1 month ago
Pass the MXC host-prep script to PowerShell as plain -Command text instead of an encoded command (#12182)
c551a10 Daniel Han committed 1 month ago
Join the npm scanner's credential path markers from pieces (#12181)
b47a0cf Daniel Han committed 1 month ago
Studio: load a downloaded GGUF from disk when the Hub cannot be read or refuses it (#12161)
866b5bc Daniel Han committed 1 month ago
Studio: run the isolated Windows Terminal on cmd.exe with stock git when Git Bash cannot start in MXC (#12164)
a19dae7 Daniel Han committed 1 month ago
Studio: tell a rejected Hugging Face token apart from an unreachable Hub (#12159)
32f806e Daniel Han committed 1 month ago
Studio: grant MXC Tier 3 read access to runtime folders once, not per launch (#12121)
1508a0e Daniel Han committed 1 month ago
Studio: fix Decision API validation and TypeSafe SDK metadata (#12186)
06911b7 Wasim Yousef Said committed 1 month ago
Studio: Triton-fused VAE norms, caches and attention for every image and video VAE (1.7x to 6.3x decode) (#12078)
e07672b Daniel Han committed 1 month ago
Read the gradient checkpointing precondition off transformers' own method (#12187)
505cfda Daniel Han committed 1 month ago
Refresh WSL shortcut icons through a Windows Python instead of an emitted P/Invoke (#12184)
407bb3b Daniel Han committed 1 month ago
Studio: pick the right model size when the file names are lowercase (#11722)
856d80e Nilay committed 1 month ago
Studio: list a crashed or cancelled run's saved checkpoints on the Export page (#11600)
7236e91 Nilay committed 1 month ago
Studio: read public Hub repos without a saved token the Hub rejects (#12158)
ecf63dd Daniel Han committed 1 month ago
Studio: list HF cache models written without symlinks in the local inventory (#12156)
6b55c9a Daniel Han committed 1 month ago