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
Tests: keep five tests off what the host has installed or downloaded (#13395)
bf10801 Daniel Han committed 1 month ago
Studio: screen a quoted program path that cmd /c or start launches (#13386)
1ecae86 Daniel Han committed 1 month ago
Studio: classify the 10 new audio.cpp-gguf folders as needing a newer runtime (#13393)
2003055 Daniel Han committed 1 month ago
Revert "Studio: keep the composer expand button in the top corner when a file is attached (#13198)" (#13394)
160cb41 Michael Han committed 1 month ago
Forward output head bias in GRPO policy scoring (#13302)
a4a8d82 cherry77-cloud committed 1 month ago
Studio: add an opt-in setting to read replies aloud automatically (#13357)
f1eb893 Daniel Han committed 1 month ago
Studio: forward llama-server /metrics, /slots, /tokenize and /detokenize (#13354)
1201351 Daniel Han committed 1 month ago
Studio: torch 2.13 for new Linux aarch64 cu130 Python 3.13 installs, existing installs keep their torch (#13380)
55ff88a Daniel Han committed 1 month ago
Studio: optional detailed Live monitor with GPU load, temperature and power (#13351)
2e38770 Daniel Han committed 1 month ago
Studio: per-account model allowlist for multi-user installs (#13359)
8bf274b Daniel Han committed 1 month ago
Studio: let LAN access bind only the addresses you choose (#12839)
af20aca Dante committed 1 month ago
Keep a user-set UNSLOTH_RETURN_LOGITS=1 across trainer.train() (#13311)
8ad2b6a Daniel Han committed 1 month ago
Studio Hub: filter models by parameter count, context length and publisher (#13363)
814c430 Daniel Han committed 1 month ago
Studio: Deep Research Retry continues with the current time budget (#13358)
e16ef2d Daniel Han committed 1 month ago
studiobench: carry on with image_upload when the click opened the menu but timed out (#13377)
501482a Daniel Han committed 1 month ago
Studio: upload files straight into a chat's folder (#13366)
658b75a Daniel Han committed 1 month ago
AsyncGRPO vLLM server: raise the listen backlog past socketserver's 5 (#13381)
489d63e Daniel Han committed 1 month ago
Studio: stop refusing C:\path\app.exe and dot-directory programs as the blocked command '.' (#13347)
6603095 Daniel Han committed 1 month ago
Studio: run React and TypeScript code blocks as live previews (#13039)
0b1291b Etherl committed 1 month ago
Train embed_tokens and lm_head in FastModel after add_new_tokens (#13057)
0ff48e8 Daniel Han committed 1 month ago
Studio: keep a wrapper-less Gemma value a string unless JSON reads it as a number (#13279)
b858369 Mohammad Hijjawi committed 1 month ago
Never train with a KV cache, so padding-free rows keep their document boundaries (#13177)
323932b Daniel Han committed 1 month ago
Fix Q-GaLore checkpoint weights and stacked MoE grouping (#13276)
0417870 cherry77-cloud committed 1 month ago
Unsloth Studio / Desktop: colour a companion download's bar by model, text encoder and VAE (#13112)
eb84cce Leo Borcherding committed 1 month ago
Studio: show the compaction count in the context meter (#13355)
c1c306a Daniel Han committed 1 month ago
Studio: bound a live Thinking block to its newest lines (#13086)
7d6354a Daniel Han committed 1 month ago
Unsloth Desktop: redact dashed sk- keys, JWTs and JSON credentials in support diagnostics (#13260)
f617649 Leo Borcherding committed 1 month ago