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: one set of sort and filter icons, shorter chat cards, toolbar tooltips (#15958)
4ff89b1 Michael Han committed 22 days ago
Studio: toast icons centred on one line, top-aligned on longer toasts (#16050)
b5052ab Michael Han committed 22 days ago
Batched MoE LoRA: every job's expert LoRA in one grouped pass on Zoo's kernels (#16012)
69117f1 Daniel Han committed 22 days ago
Studio MCP: bound get_job wait_s with a fixed limit, not the patchable MAX_WAIT_S (#16033)
9d3caca Daniel Han committed 22 days ago
Studio: int8 Qwen-Image text stream when the denoiser is streamed (1.24x under a 16 GiB budget) (#16027)
aca386f Daniel Han committed 22 days ago
Studio video: streamed mp4 byte-identical to the stock export on every render (cap libx264 at AVX2 on AVX-512 hosts) (#16030)
5a651b1 Daniel Han committed 22 days ago
Studio video: compare streamed frames without chunk-sized temporaries (H3 export stalls) (#16024)
d84e074 Daniel Han committed 22 days ago
Studio: Neural Engine Prefill toggle for MLX models (#15954)
70750d9 Long Yixing committed 22 days ago
Studio: hand the audio route's heal gate the request's contract, and follow the hoisted audio_base64 in the turn-scope check (#16023)
8c49443 Daniel Han committed 22 days ago
Tests: read the training API job bodies fixture as UTF-8 (#16025)
61176d5 Daniel Han committed 22 days ago
T4: hand batch-4 NF4 decode to the bitsandbytes gemm_4bit where it runs (#15988)
c97cd18 Daniel Han committed 22 days ago
Studio: say why AI Assist cannot map a dataset with no prompt/response split (#15991)
242812e Daniel Han committed 22 days ago
Studio: batch MLX tool-protocol, response_format and video replies; several clips and Gemma 4 video on MLX (#15920)
cb07e19 Long Yixing committed 22 days ago
Studio video: request a hosted .safetensors twin before each declared prequant .pt (#15980)
8b17010 Daniel Han committed 22 days ago
Studio diffusion: Qwen-Image text-to-image activation seeds keep part of the int8 DiT resident under a VRAM budget (#15985)
2aedb67 Daniel Han committed 22 days ago
Collapse blank line runs with a literal-led pattern (about 10x faster) (#15950)
761487f Daniel Han committed 22 days ago
Tests: price the auto-context fit floor as the layer-split path on any host (#16014)
8ecab3d Daniel Han committed 22 days ago
Move the Unsloth Studio MLX pins to mlx-vlm 0.7.11 (#15861)
3cc31f7 Long Yixing committed 22 days ago
Studio: load the chat page on first use instead of in the startup bundle (#15943)
c45bfe5 Daniel Han committed 22 days ago
Tests: keep the RAG query-packing test on vec0 while the in-memory cache builds in the background (#16013)
86be805 Daniel Han committed 22 days ago
fix(studio): attach an explicitly chosen MLX drafter even when it is priced not to fit (#15813)
316e575 Long Yixing committed 22 days ago
Unsloth Studio: LFM3-Audio / LFM3-Omni on MLX, with spoken replies (#15889)
d7aa02b Long Yixing committed 22 days ago
Studio: train single-string-column datasets as raw text (#15899)
ffd1a18 Yakshith committed 22 days ago
Studio diffusion: use an installed HIP SDK for the Windows ROCm sd.cpp build when its bin dir is not on PATH (#15853)
35492de oobabooga committed 22 days ago
Studio setup: skip the OXC validator npm ci when its installed tree is unchanged (#15947)
d692510 Daniel Han committed 22 days ago
Studio video: stage an fp32 encoder on the card when its dense cast would not fit host RAM (#15975)
d3702be Daniel Han committed 22 days ago
Studio: stop the Claude CLI billing line from breaking prompt caching (#15845)
453b788 Nilay committed 22 days ago
Studio: recommend a Hub GGUF quant that fits the Mac's Metal budget (#15848)
0f51e21 Nilay committed 22 days ago
Studio: reuse lexical passes and branch checks across a compaction and prune branch-run starts (#15972)
0d9b63e Daniel Han committed 22 days ago