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: use the Hugeicons internet icon for every globe (#12321)
5a45952 Michael Han committed 1 month ago
Studio desktop: Cmd/Ctrl + and - zoom with a zoom popup (#12280)
b8440cc Michael Han committed 1 month ago
Studio: one top row and layout fixes for the Windows desktop app (#12070)
c309bd6 Maheswar Kumar committed 1 month ago
Unsloth Studio / Desktop: expose prefill progress in API monitor (#11161)
5651def Maheswar Kumar committed 1 month ago
Studio: stop a generalized compare cleanly, cancelling its in-flight model load (#7407)
a790248 Lee Jackson committed 1 month ago
Studio: run large Laya requests in token-budgeted chunks (#12271)
1bf9702 Daniel Han committed 1 month ago
Studio: list large MCP tool schemas compactly and load the full schema on demand (#11046)
fb5a921 Nilay committed 1 month ago
feat(studio): Mac adapter-format option for LoRA export, with GGUF adapters on macOS (#7539)
46161d7 Long Yixing committed 1 month ago
Preserve repo id case when resolving a model name (#8058)
aa36c8d BardiaKoopah committed 1 month ago
feat(studio): reuse MLX VLM prompt snapshots in the resident vision batch (#12263)
9bafddb Long Yixing committed 1 month ago
Train gpt-oss MXFP4 LoRA through unsloth-zoo's packed experts when load_in_16bit is not set (#11929)
b7a0d71 Daniel Han committed 1 month ago
Keep LoRA on dense Linears sharing a name with fused MoE experts in PEFT's v5 conversion (#12281)
7eb3de3 Daniel Han committed 1 month ago
Let FastModel and FastLanguageModel load the same family in one process, in either order (#12219)
d2e127d Daniel Han committed 1 month ago
Harden timing-dependent Studio backend tests (nvidia-smi cache, cold /api/health) (#12275)
311572e Daniel Han committed 1 month ago
Unsloth Studio / Desktop: reload llama.cpp models on reconnect (#11341)
c72fa04 Maheswar Kumar committed 1 month ago
Fix: Support past_key_values in model.generate for multi-turn conversations (#4232)
0dae825 Daniel Han committed 1 month ago
Unsloth Studio installer (AMD/Windows): install PyTorch for RX 5000 to 9000 cards from AMD's multi-arch index (#11846)
f00582a Leo Borcherding committed 1 month ago
Baseline the thirteen unsloth-zoo 2026.9.8 findings after review (#12292)
c4a8135 Daniel Han committed 1 month ago
Studio: keep unsloth start alive while a LoRA adapter's base model downloads (#10550)
fd1ddeb Nilay committed 1 month ago
Keep a model's no-placement parameters on CPU (Qwen3.8-Flash-Next n-gram table) (#12141)
431ce2d Daniel Han committed 1 month ago
Unsloth Studio / Desktop: support custom vision projectors (#11352)
6c1cc50 Maheswar Kumar committed 1 month ago
Unsloth Studio / Desktop: show the GPU llama.cpp runs on when it is CUDA and differs from training (#12251)
a521448 Leo Borcherding committed 1 month ago