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: resolve the FLUX.2-klein 9B base for a local ComfyUI file (#13290)
aced791 Daniel Han committed 1 month ago
Studio: let an MCP call run again after other work changed state (#13287)
0006f0e Daniel Han committed 1 month ago
Studio: keep a media chat's context usage after a local model switch (#13281)
6ad3fa6 Daniel Han committed 1 month ago
Studio (Jetson): size llama-server context from system RAM when nvidia-smi reports [N/A] (#13286)
b2925e7 Daniel Han committed 1 month ago
Studio: keep auto-load and compare panes off the default engine for compressed-tensors, AWQ and GPTQ checkpoints (#13229)
86c5ac2 Daniel Han committed 1 month ago
Studio: start vLLM on hybrid models whose Mamba state outruns max_num_seqs (#13235)
f64ed0d Daniel Han committed 1 month ago
Installer: name the missing tools, not Debian packages, on non-apt Linux (#13284)
5b4a872 Daniel Han committed 1 month ago
Studio: name a missing GGUF metadata key instead of blaming the file or memory (#13285)
62cbee3 Daniel Han committed 1 month ago
Studio: open Images workflows from More when Images is unpinned (superseded by #13101) (#12767)
6954690 Abir Das committed 1 month ago
Studio: check loopback for an existing listener before taking the port (#11634)
e2bd4d2 Yeluru Nitish committed 1 month ago
Studio: API monitor Load button (superseded by #11223) (#10378)
ac5a1bb michupeel committed 1 month ago
Studio: stop charging SWA context checkpoints against VRAM on discrete GPUs (#13242)
967dcda Daniel Han committed 1 month ago
Run the RoPE and NF4 Triton kernels on GPUs older than Volta (#13153)
27e088f Daniel Han committed 1 month ago
Treat Docker on WSL2 as WSL when deciding embedding offload (#13259)
3e9076d Daniel Han committed 1 month ago
Pass GradScaler to GKD chunked JSD with an older-Zoo fallback (#13190)
8957742 cherry77-cloud committed 1 month ago
Skip the fused LoRA kernels when a DoRA adapter is active (#13254)
2308f6e Vineeth Sai Varikuntla committed 1 month ago
Keep the padding mask for left-padded label-less training forwards (#13154)
62354ed Daniel Han committed 1 month ago
Studio vLLM: allow Triton non-constexpr globals so DeepSeek-OCR loads (#13264)
65e2314 Daniel Han committed 1 month ago
Packed SDPA rows: one torch varlen_attn call instead of one SDPA call per segment (#13221)
2ea28dd Daniel Han committed 1 month ago
Fix fp16 LoRA training with plain transformers.Trainer (fused CE gradients underflowed) (#13152)
f951515 Daniel Han committed 1 month ago
Studio: pick up a GPU whose driver loads after Studio starts (#13232)
c0f64cd Daniel Han committed 1 month ago
Fix steering follow-ups disappearing when a response is cancelled (#13249)
028db7b Lee Jackson committed 1 month ago
Studio: title refresh follows the connection that answered the chat (#13274)
2a9f621 Daniel Han committed 1 month ago
Studio: confine native dataset drops to the account workspace; cap child API key expiry (#13273)
a7bddd7 Daniel Han committed 1 month ago
Docs: secure-mode Docker comment also binds JupyterLab to 127.0.0.1 (#13275)
bf0d467 Daniel Han committed 1 month ago
Studio: select all and shift-click ranges in the Library (#13250)
b1f4fdb Padias committed 1 month ago
chore(deps): apply six validated Studio dependency bumps (#13148)
38c4f67 Wasim Yousef Said committed 1 month ago
OpenClaw 2026.10.1: define the CI agent under keyed agents.entries (#13266)
beb556e Daniel Han committed 1 month ago
Disconnect tests: give the hanging client seconds, not 0.2 s, to be reached (#13272)
c249dca Daniel Han committed 1 month ago
Studio installer: drop a flash-attn built for a torch that is gone, and pair torch 2.9 with torchao 0.15.0 (#13247)
6d7c6e9 alkinun committed 1 month ago