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: finite MPS memory cap and sd.cpp --mmap on macOS (#13763)
35c74c8 Daniel Han committed 1 month ago
Studio: show the browser agent cursor where it clicks and types, with a Settings toggle (#13729)
52a3fbe Michael Han committed 1 month ago
Studio: keep list numbers in Word and HTML knowledge base files (#13755)
1a76e88 Nilay committed 1 month ago
Studio: let tool results use the window when auto-compaction is off (#13811)
4dc3f93 Daniel Han committed 1 month ago
Studio: use native Linux browser tabs with GTK overlay sizing (#13717)
02d9ea3 Wasim Yousef Said committed 1 month ago
Studio: train ShareGPT tool calls natively on processor-loaded models (#13824)
361e825 Daniel Han committed 1 month ago
Studio: tell the model when a site's bot protection blocks a page fetch (#13760)
42735fc Nilay committed 1 month ago
Score a 16-bit fine-tune with unsloth eval in 16-bit, not in 4-bit (#13758)
2a8def3 Nilay committed 1 month ago
Studio: tighter hover pill on the chat model selector (#13817)
2e3d773 Michael Han committed 1 month ago
Studio: install the Vulkan or CPU llama.cpp bundle on Windows NVIDIA drivers older than CUDA 12 (#13671)
c155cd7 oobabooga committed 1 month ago
Studio: keep a Connection's thinking between tool calls in the same answer (#13785)
77252b3 oobabooga committed 1 month ago
Studio: each chat keeps its own browser tabs, and the browser agent works on in chats off screen (#13800)
e16e7fc Daniel Han committed 1 month ago
Studio: read the code behind a GitHub file link in web search (#13754)
eff9bb4 Nilay committed 1 month ago
unsloth start opencode: never pass --model on a bare launch; OpenCode 2 rejects it (#13783)
8060a5e Yogesh Seenichamy committed 1 month ago
Studio: let YuE2 songs run to their own ending instead of 1.5x the length (#13371)
1d79d47 Daniel Han committed 1 month ago
Pin vermin: 1.9.0 fails the backend floor lint on valid 3.5 syntax (#13818)
8116dd2 Daniel Han committed 1 month ago
Unsloth Studio / Desktop (AMD, Windows): install the Vulkan llama.cpp prebuilt when device 0 has no HIP bundle (#13680)
ccae5d4 Leo Borcherding committed 1 month ago
Studio: unload a model under the id the API lists for it (#13810)
9396996 Daniel Han committed 1 month ago
Studio: stopping a model load leaves no Starting model status behind (#13803)
82c608b Michael Han committed 1 month ago
Give the data settings driver's cold first load the same 120s as its mount wait (#13809)
fd8d3bb Daniel Han committed 1 month ago
Build Windows x86_64 prebuilt CUDA wheels for torch 2.12 (#13807)
d6fc96b Daniel Han committed 1 month ago
Studio: name the Windows update that adds MXC's built-in container (#13798)
2447d83 Daniel Han committed 1 month ago
Measure the resume glyph's centring in viewBox units, not painted rects (#13799)
3c0061e Daniel Han committed 1 month ago
Studio: Windows torch 2.10 kernels from our prebuilt wheels (#13742)
6aa41b5 Daniel Han committed 1 month ago
Studio: cache Edit-mode text encoding by reference image content (#13683)
a858567 kevin committed 1 month ago
Studio: keep the desktop browser page live when it opens as a dialog (#13734)
8756dfd Daniel Han committed 1 month ago
Studio: raise the startup bundle budget by #13589's measured growth (#13789)
6c3bd16 Daniel Han committed 1 month ago