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: stop rejecting --mmproj-device CUDA1 when gpu_ids are saved (#12505)
a91ee37 Stefan Wang committed 1 month ago
Studio: renew the chat-run lease while long prefill is still advancing (#12523)
d8740d3 Indrajit swain committed 1 month ago
Count only the lock's own waits in the unlockable-filesystem row (#12529)
5b78baf Shubhankar Tripathy committed 1 month ago
Studio: move torchao int8 weights back to the GPU after an oversized request streams pinned groups (#12526)
d883d2d Daniel Han committed 1 month ago
Studio: use ComfyUI's default settings for FLUX.1, Qwen-Image, Z-Image, Ideogram 4, Wan2.2 and HunyuanVideo-1.5 (#12516)
fd2522b Daniel Han committed 1 month ago
Parity: keep the reap-margin check off the edge of its own window (#12514)
dae96ea Shubhankar Tripathy committed 1 month ago
Studio: keep the composer permission shield still while the plus spins (#12506)
77b0296 Michael Han committed 1 month ago
Studio: clip every rounded scroller in Firefox, dropping the overflow watcher (#12438)
2c72350 Michael Han committed 1 month ago
Studio: center the sidebar row kebab and pin in their hover circle (#12502)
11cc2ab Michael Han committed 1 month ago
Studio: draw the command palette like chat search (#12511)
767a9f0 Michael Han committed 1 month ago
Studio: add audio.cpp as a native engine for speech, music and dictation (#12342)
cd9162f Etherl committed 1 month ago
Studio: run vLLM and SGLang on Windows inside a private WSL2 distro (#12024)
619cd71 Daniel Han committed 1 month ago
Lift the transformers ceiling to 5.17.0 and align the mirrored CI caps (#11016)
9f85f53 Daniel Han committed 1 month ago
Studio: add opt-in vLLM and SGLang support with multi-GPU inference, quantization and vision (#11491)
08716f0 oobabooga committed 1 month ago
Make RMSNorm, RoPE and the input-embedding hook traceable by torch.compile (#12171)
5d75cde Daniel Han committed 1 month ago
Fused Triton NF4 dequant and GEMV: bit-exact, stream-safe, torch.compile traceable (#12113)
338d23f Daniel Han committed 1 month ago
Studio: add per-model custom llama.cpp INI configuration (#10783)
864fec4 Etherl committed 1 month ago
Studio: train a dataset's system column as the system prompt (#12486)
ed6d0e8 Nilay committed 1 month ago
Studio: stop CLI commands running as another account when Studio has more than one (#12492)
6f0278d Nilay committed 1 month ago
Log the real reason the Decision API model failed to load (#12491)
c6363d3 Nilay committed 1 month ago
Studio: load the cached int8 checkpoint for a 5-bit-or-below GGUF pick that has to offload (Qwen-Image-2.1 about 2x faster at 12 / 8 GB) (#12455)
f2f09ec Daniel Han committed 1 month ago
Studio: keep MiniMax-H3 GGUF renders on a resident sd-server, released when idle (#12461)
c29edef Daniel Han committed 1 month ago
Studio: keep Word equations when a .docx is added to a knowledge base or chat documents (#12490)
7d21673 Nilay committed 1 month ago
Studio: keep <placeholder> and Vec<T> text in research reports and file previews (#12489)
47ecafb Nilay committed 1 month ago
Studio: keep tool-call arguments when training on tool-calling datasets (#12487)
666ad7b Nilay committed 1 month ago
Studio: stop adding backslashes to Word files in Data Recipes (#12488)
48075b4 Nilay committed 1 month ago
Studio: make unsloth start sampling and reasoning flags apply only to that agent (#12493)
016a2b5 Nilay committed 1 month ago
Studio: faster first renders and resident text encoder under group offload (#12519)
e700a33 Daniel Han committed 1 month ago
Studio: run Z-Image and Wan2.2-5B / HunyuanVideo-1.5 in fp16 on fp16-only GPUs instead of fp32 (#12476)
77c1934 Daniel Han committed 1 month ago
Studio: load FLUX.1 and Qwen-Image on fp16-only GPUs with small host RAM (#12480)
0a0da52 Daniel Han committed 1 month ago