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: show an error when Gemini fails a reply partway through (#12242)
4a77419 Nilay committed 1 month ago
Studio: fit the context again when MTP is forced on a reload (#12231)
f160cf5 Nilay committed 1 month ago
Studio: title a chat with the connection that answered it, not the local model (#9106)
1bfd123 Long Yixing committed 1 month ago
Unsloth Studio: explain why Model Memory skips the RAM lock when a model is fully on the GPU (#9571)
03ecc3b Daniel Han committed 1 month ago
Scale the attachment preview text with the UI font size (#12276)
6200666 Daniel Han committed 1 month ago
Wait for the model-config Load to finish before the test moves on (#12270)
21a835f Daniel Han committed 1 month ago
Studio: do not fail a cached dataset load on a bookkeeping write (#8057)
7617fc3 BardiaKoopah committed 1 month ago
Add Qwen2.5 Coder model support to registry and chat templates (#4221)
fa61ee4 Daniel Han committed 1 month ago
Average GRPO gradients across DDP ranks, and keep MiniCPM3's pre-head scaling (#12193)
1c9bf9f Daniel Han committed 1 month ago
Read the llama.cpp changelog after it settles, by content rather than innerText (#12273)
b2b813f Daniel Han committed 1 month ago
Studio: let the chat model use the Decision API through MCP (#12232)
e3dced1 Nilay committed 1 month ago
Reuse the dY @ B.t() product in the fused LoRA backward (#12212)
ae7ed40 Daniel Han committed 1 month ago
Fall back to PEFT's LoRA forward for float32 base weights (#7020)
5a60350 Daniel Han committed 1 month ago
Keep import unsloth working when vLLM needs transformers 5 (#12268)
62db978 Daniel Han committed 1 month ago
Composer attachments: first page previews, Library icons, lighter cards (#12261)
6fabaab Michael Han committed 1 month ago
Studio: read a remote image URL on safetensors and MLX models instead of ignoring it (#12243)
1783a35 Nilay committed 1 month ago
Studio: stop tool calls from getting stuck in Running forever (#12234)
2b76e7d Nilay committed 1 month ago
Studio: stop chats with images from blocking other chats on GGUF models (#12236)
6aac263 Nilay committed 1 month ago
Fix NaN gradients from compiled flex attention on short static training batches (#12207)
e22e297 Daniel Han committed 1 month ago
Studio: name the llama.cpp graph-scheduler abort and stop reloading into it (#6741)
9a2c982 Daniel Han committed 1 month ago
Studio: push only the exported merged model to the Hub on a Mac (#12239)
852795d Nilay committed 1 month ago
Studio: say when a model only partly fits on the GPU (#8812)
01b8b84 Daniel Han committed 1 month ago
Run the user message timestamp Playwright check in Frontend CI (#12255)
529a6a7 Daniel Han committed 1 month ago
Library: equal size chat cards and no loading placeholders (#12221)
6c20f0f Michael Han committed 1 month ago
Studio: faster Laya inference (marker-only head, CUDA graphs) (#12224)
4e8503e Daniel Han committed 1 month ago