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: fix Base vs LoRA compare for voice messages on fine-tuned Whisper (#12381)
a6bf9b3 Nilay committed 1 month ago
Studio: train an uploaded CSV's NA, None and 00501 cells as written (#12379)
14ebbe8 Nilay committed 1 month ago
Only turn on HF_HUB_ENABLE_HF_TRANSFER in synthetic.py when hf_transfer is installed (#12406)
4004bf7 Daniel Han committed 1 month ago
Create the SentencePiece scratch directory without a check-then-create race (#12400)
8568ae3 Daniel Han committed 1 month ago
Read a sign-flipped SVD basis as the same subspace in the Q-GaLore schedule (#12399)
c4d2e24 Daniel Han committed 1 month ago
truststore: keep TLS verification on when handshakes overlap on one context (#12403)
0840364 Daniel Han committed 1 month ago
Remove xFormers built for another torch after Linux repair (#11783)
0dd7dfc xiao committed 1 month ago
Harden installer source selection, ROCm helper staging and npm scanner cleanup (#12402)
953f99f Daniel Han committed 1 month ago
Studio: keep <placeholder> and Vec<T> text in chat replies (#12376)
28d8a04 Nilay committed 1 month ago
Studio: count companion assets in the Hub download size of image and video GGUFs (#12370)
a76ac47 oobabooga committed 1 month ago
Studio: stop offering Claude sampling settings that are silently ignored (#12383)
f8eb9ba Nilay committed 1 month ago
Studio: keep the answers typed into a PDF form when it is attached to a chat (#12377)
5eccfd9 Nilay committed 1 month ago
fix(device_type): flush and fence through the shared device helpers (#12284)
ed7a078 li-lizhe committed 1 month ago
Studio: compact context usage ring when the chat header is squeezed (#12355)
b96af04 Michael Han committed 1 month ago
Studio: keep the model name ahead of its format and quant when the header is tight (#12396)
ffd71a9 Michael Han committed 1 month ago
Studio: keep int8 / fp8 image transformers quantised when the load has to offload (#12287)
e477bad Daniel Han committed 1 month ago
Baseline the moved transformers 5.18.0 testing_utils polling-loop site after review (#12368)
fa39581 Daniel Han committed 1 month ago
Studio frontend test: drain the previous runtime's writes before each reasoning-effort scenario (#12367)
ac0ce6f Daniel Han committed 1 month ago
Studio: keep INT8 and compile on conventional video models when they have to offload (#12289)
e8b2bb2 Daniel Han committed 1 month ago
Model-config UI test: stop waiting on a request whose finish event never arrives (#12363)
5caccb9 Daniel Han committed 1 month ago
Studio: keep MiniMax-H3's int8 denoiser and compile when it has to offload (#12286)
5879a9b Daniel Han committed 1 month ago
Studio video: keep an explicit fast LTX-2.3 single-file load resident when it fits (#12345)
ba7ff6e Daniel Han committed 1 month ago
Build the marlin_gemm call from the op schema so packed INT4 inference works on vLLM 0.29 (#12320)
7740a23 Daniel Han committed 1 month ago
Studio: load downloaded tokenizers offline, size unsloth mirrors from the family table (#12300)
7cee0c1 Daniel Han committed 1 month ago
Keep every active LoRA adapter in the fused LoRA paths (#12214)
b70bd28 cherry77-cloud committed 1 month ago
Unsloth Studio (AMD): run native image generation on an NVIDIA card next to ROCm torch (#12252)
1a08bd9 Leo Borcherding committed 1 month ago