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: do not offer a local model that is short a shard (#8985)
88cfd06 sts-change committed 1 month ago
Studio: load the ComfyUI-format twin of a hosted image prequant, and run ComfyUI fp8 on the fp8 path (#12820)
b4aa7b1 Daniel Han committed 1 month ago
Studio: Sandbox Low/High and Permissions under Settings > Sandbox (#12802)
bcdede6 Daniel Han committed 1 month ago
Seq2Seq LoRA task type and GA token count for T5Gemma2 (#7795)
b5fd144 Daru Okta Buana committed 1 month ago
Studio: share one kernel install core with unsloth install-kernels (#12828)
2622a5d Daniel Han committed 1 month ago
Studio: run Qwen-Image-2.1 edits on 16 GB cards instead of refusing them (#12752)
dfc56d9 oobabooga committed 1 month ago
Wait on the pooled fetch itself in the closed-request browser test (#12835)
42719e4 Daniel Han committed 1 month ago
Studio: resolve UUID-form CUDA_VISIBLE_DEVICES to physical GPU indices (#8917)
5169487 Darshan Poudel committed 1 month ago
Keep EmbeddingGemma and Qwen3-Embedding prompts when fine-tuning (#12795)
2f55d63 Nilay committed 1 month ago
Studio: open chat HTML in the Browser and remove Canvas (#12804)
6a1ca43 Michael Han committed 1 month ago
Clear the duplicate-definition backlog and gate it at zero (#9769)
b349936 Rajarshi Datta committed 1 month ago
Studio: load ComfyUI-format int8_convrot / fp8 single-file DiTs instead of rendering noise (#12765)
1d9accd Daniel Han committed 1 month ago
Studio: build each Qwen-Image-Edit variant on its own transformer config (2509 / original Edit no longer get 2511's zero_cond_t) (#12815)
4b00d49 Daniel Han committed 1 month ago
Studio: make the OS sandbox work in Colab and other containers (#12801)
813525e Daniel Han committed 1 month ago
feat(studio): opt-in reasoning for Custom Chat Completions connections (#12743)
6530c3a Wasim Yousef Said committed 1 month ago
Studio: keep a Library file's star, folder and name after an edit (#12793)
c120fc8 Nilay committed 1 month ago
Fix MXC inherited runtime permissions and pending grant recovery (#12758)
8a469fe Lee Jackson committed 1 month ago
Fix Windows MXC runtime grants for uv-managed Python (#12756)
003448a Lee Jackson committed 1 month ago
Studio: add model picker and reload shortcut to API monitor (#11223)
769b34f Ayushman committed 1 month ago
Studio: add custom validator blocks to Data Recipes (#7784)
010aac8 Kent Kawashima committed 1 month ago
Studio: let safetensors vision models see every image in a chat (#12787)
ce74c2e Nilay committed 1 month ago
fast_inference for Qwen3.5 / 3.6 MoE and Gemma-4 MoE with LoRA on the experts (#12742)
f8f7467 Daniel Han committed 1 month ago
Studio: dismiss the llama.cpp update toast once the job finishes (#9199)
9df22eb Indrajit swain committed 1 month ago
Run narrow RMSNorm rows several per program, bit-identical (Qwen3 q/k norms) (#12754)
b945fd7 Daniel Han committed 1 month ago
Studio: dark dropdown glow, menu ticks and picker polish (#12826)
a01c0f8 Michael Han committed 1 month ago