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 Hub: rank Latest by trending against creation date, not last modified (#14329)
ad2fc06 Michael Han committed 1 month ago
Studio: suggest Unsloth Docs in the web UI browser panel (#14324)
3c68c2a Michael Han committed 1 month ago
Native FSDP2: fully_shard per block and root for every model, LoRA, QLoRA and full finetuning (UNSLOTH_FSDP2=1) (#14273)
c280b6f Daniel Han committed 1 month ago
Studio: fix Claude API reference pages in the browser panel (#14323)
c87f845 Wasim Yousef Said committed 1 month ago
Studio: load ComfyUI fp8mixed text encoders (nvfp4 / mxfp8 layers) (#14321)
b6effa3 Daniel Han committed 1 month ago
Studio: open the Unsloth docs and Next.js sites in the browser panel (#14279)
eb0a983 Nilay committed 1 month ago
Raise the startup bundle budget for the FLUX.3 controls (#14312) (#14319)
360cd9a Daniel Han committed 1 month ago
SAM 3 skeleton test: compare against a float32 load and find the RoPE tables by any spelling (#14318)
5c8ccce Daniel Han committed 1 month ago
Model catalog network check: judge FLUX.3 Klein by the mirror the backend fetches (#14316)
f142f57 Daniel Han committed 1 month ago
Studio: fix FLUX.3 workflows, controls, LoRAs and download progress (#14312)
b32d544 oobabooga committed 1 month ago
Studio: logprobs on /v1/completions and /tokenize for MLX models (#14229)
6a19cb2 Daniel Han committed 1 month ago
Studio: serve /v1/completions from managed vLLM and SGLang engines (#14227)
1935c8a Daniel Han committed 1 month ago
Studio: SAM 3.1 for Mask from text, with pre-quantized INT8 / FP8 files (#14310)
fb372c1 Daniel Han committed 1 month ago
Baseline the unsloth-zoo 2026.10.7 scan findings after review (#14311)
e326020 Daniel Han committed 1 month ago
Studio: Library chat cards keep their column width and equal padding (#14305)
22a0bea Michael Han committed 1 month ago
whisper.cpp: skip the macOS release walk when llama.cpp is ahead of every release (#14302)
87b8c45 Daniel Han committed 1 month ago
Import fix: torchao 0.17+ on torch 2.6 (torch.utils._pytree.register_constant) (#14299)
2c32459 Daniel Han committed 1 month ago
Studio: update external llama.cpp default port to 9931 (#14266)
56d6f75 oobabooga committed 1 month ago
Avoid unnecessary attachment reindexing on chat generation (#14217)
30ca824 Lee Jackson committed 1 month ago
fix(rag): avoid long CPU stalls when archiving tool calls (#14098)
104bc89 Dmitry committed 1 month ago
Retry the audio.cpp server on a new port when its port is taken (#14294)
9adeecf Daniel Han committed 1 month ago
GRPO: reduce the IS ratio metric per sequence for geometric modes and the env override (#14192)
b6f21fd Daniel Han committed 1 month ago
FSDP2 QLoRA: keep packed 4-bit bytes out of the loader dtype pass, fix merged_4bit save (#14165)
37745d5 Daniel Han committed 1 month ago
Studio: change Parallel Slots from the API page's Slots readout (#14225)
cd2595f Daniel Han committed 1 month ago
CI: raise the Core zoo job cap from 25 to 35 minutes (#14240)
4f5a98c Daniel Han committed 1 month ago
Review follow-ups: fused LoRA dropout/bias under compile and 4-bit, Q-GaLore offload, GRPO truncated support, CP save on transformers 4.x (#14226)
103ba52 Daniel Han committed 1 month ago
Optimizer parking outside inference mode; refuse fp16 params with unquantized adamw_flash moments (#14201)
143e41f Daniel Han committed 1 month ago
Large head attention follow-ups: compiled kill switch, 4-bit shard self check, forced float32 skip (#14188)
73d2598 Daniel Han committed 1 month ago