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: warm the compile workers and Triton / inductor keys during a load that compiles inline (#16918)
1624f01 Daniel Han committed 14 days ago
Studio: restore WebKitGTK offscreen code fences in batches and on any pointer release (#16919)
07a9c87 Daniel Han committed 14 days ago
Studio: merge LoRA and LoKR adapters into ComfyUI-quantized weights (#16874)
61f2c06 oobabooga committed 14 days ago
Studio GGUF: keep the stock Q8_0 forward on torch below 2.7 (#16911)
257e925 Daniel Han committed 14 days ago
Bump install.sh / install.ps1 pins to unsloth>=2026.11.3 (#16920)
bccbeca Daniel Han committed 14 days ago
Studio: reduce long-chat panel and scrolling lag on Linux (#16854)
6c2b895 Wasim Yousef Said committed 14 days ago
Studio: stop dialogs and menus from restyling the whole chat on open and close (#16902)
3738b40 Daniel Han committed 14 days ago
Reduce scrolling lag in long chats by bounding offscreen reply painting (#16782)
a17436b Lee Jackson committed 14 days ago
Studio: Element outline setting, outlined light glass, Transparent glass with strength as opacity, glass Show on, and glass on pages without a background (#16890)
dc916c9 Michael Han committed 14 days ago
Studio: wait for a pending embedding model download instead of failing every linked-folder file (#16722)
2ea8662 Debojyoti Banerjee committed 14 days ago
Studio: keep fine-tuning job pages complete when a batch starts (#16823)
2e10d80 Nilay committed 14 days ago
fix(studio): preserve Unicode separators in S3 JSON Lines transcripts (#16725)
1b20b9e Max-Reisinger committed 14 days ago
Chat UI: let the floating monitor settle after the viewport restore before dragging it (#16909)
b9f8341 Daniel Han committed 14 days ago
Tests: undo the Unsloth load's PEFT LoRA forward swap after the probe memo tests (#16908)
7a13256 Daniel Han committed 14 days ago
Studio: fill the H3 conditioner's pinned arenas once, before registering (33% faster cold pin) (#16887)
ccc8a82 Daniel Han committed 14 days ago
Studio: prequant fingerprint check on a thread pool, zero-copy (13x faster first load verify) (#16879)
2a6d891 Daniel Han committed 14 days ago
Studio LTX-2.5: read the hosted int8 checkpoint as native Linears when the DiT is offloaded (#16886)
250895c Daniel Han committed 14 days ago
Studio diffusion: overlap the quant scheme probe with the planner import on a cold image load (#16882)
9c08475 Daniel Han committed 14 days ago
Studio: fix the browser use bar's text size and button order (#16905)
f85c190 Michael Han committed 14 days ago
Studio: smaller sidebar logo on web, desktop sidebar header back to its pre-#16775 look (#16907)
b3cccf3 Michael Han committed 14 days ago
Document browser use and sandboxing in the README (#16903)
04a0fd7 Lee Jackson committed 14 days ago
Studio: reorder cloud providers in the connection picker (#16904)
f957c78 Michael Han committed 14 days ago
Studio Qwen-Image-2.1: fused q/k RMSNorm + RoPE and broadcast modulation for eager renders on ROCm (#16891)
a6cd920 Daniel Han committed 14 days ago
Studio diffusion: settle device memory once per video plan, wake sd.cpp polls on server output (#16797)
416cedd Daniel Han committed 14 days ago
Keep gradient checkpointing settings working on transformers 5.23 (#16833)
e8aaca2 Nilay committed 14 days ago
Backend tests: importorskip the unsloth submodule a test uses, not the stubbable package (#16898)
2bfb197 Daniel Han committed 14 days ago
Studio: match the Chat with files panel and Add files button to the glass theme (#16761)
3784089 Lee Jackson committed 14 days ago