vrr/unsloth
2
0
Fork 0
mirror of https://github.com/unslothai/unsloth.git synced 2026-07-18 04:03:49 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 25

Commit graph

  • 74917b6bb0 Move the CPU retry out of the exception handler studio-stt-dictation shimmyshimmer 2026-07-17 20:56:25 -07:00
  • 400c950eee Fix video progress under-reporting during load and generate image-generation Daniel Han 2026-07-18 03:48:54 +00:00
  • db72016c39 Fix silent LoRA drop and wasted transformer prefetch on GGUF quant loads video-diffusion-improvements Daniel Han 2026-07-18 03:49:05 +00:00
  • 0d866998f0 Fix silent LoRA drop and wasted transformer prefetch on GGUF quant loads Daniel Han 2026-07-18 03:49:05 +00:00
  • 914381ee01 Fix video progress under-reporting during load and generate Daniel Han 2026-07-18 03:48:54 +00:00
  • ec27dea236 Tighten the blocked-watchdog stub comment agent/stabilize-studio-regression-tests Daniel Han 2026-07-18 02:15:17 +00:00
  • 361cbbb287 Wire the hosted HunyuanImage 2.1 int8/fp8 checkpoints Daniel Han 2026-07-17 23:59:07 +00:00
  • dc3e079318 Wire the hosted HunyuanImage 2.1 int8/fp8 checkpoints Daniel Han 2026-07-17 23:59:07 +00:00
  • 685f843696 Stabilize Studio regression tests Daniel Han 2026-07-17 23:54:29 +00:00
  • a8ff8673df
         feat(studio): expose an opt-in MCP control plane (#7191) main Ritwij Aryan Parmar 2026-07-17 19:50:39 -04:00
  • a14b032d79
         Propagate fp8 block_size before the early return in get_lora_parameters_bias (#7189) Vineeth Sai 2026-07-17 16:30:45 -07:00
  • bf4185a2d3
         Studio: don't apply nest_asyncio on plain CLI starts (breaks asyncio on Python 3.14+) (#7186) Nilay 2026-07-18 04:57:28 +05:30
  • 8ff2f8e70c
         fix(studio): ignore reasoning in tool reprompts (#7134) Long Yixing 2026-07-18 07:22:11 +08:00
  • 53836cc62e Never let legacy migration evict existing per-model configs studio-picker-followup danielhanchen 2026-07-17 23:18:52 +00:00
  • 1b3a47b626 Accept the HF token from the header or the legacy query param danielhanchen 2026-07-17 23:18:52 +00:00
  • 1b68c2c0a0 Pin the measured HiDream quant verdict in tests Daniel Han 2026-07-17 13:08:53 +00:00
  • 958645bb6b Pin the measured HiDream quant verdict in tests Daniel Han 2026-07-17 13:08:53 +00:00
  • 5441266e3a
         Fix Studio reasoning channel rendering (#7121) Long Yixing 2026-07-18 06:30:37 +08:00
  • 2139200b3f
         Studio: don't re-download updated GGUFs on load (#7209) oobabooga 2026-07-17 19:05:46 -03:00
  • 57785f92d2
         fix(mlx): relax context-store timeout by default (#7141) Long Yixing 2026-07-18 04:13:19 +08:00
  • 49f2879cf8
         fix(studio): recover stalled Hub downloads over HTTP (#6858) Rod Boev 2026-07-17 12:32:03 -04:00
  • 069aabb564 Fix stale sidebar account-block test regex after markup change fix-sidebar-leading-tight-regex danielhanchen 2026-07-17 12:57:17 +00:00
  • 7b7c9bf898 [pre-commit.ci] auto fixes from pre-commit.com hooks pre-commit-ci[bot] 2026-07-17 14:40:22 +00:00
  • 606b7c86a5 Address review feedback on STT snapshot checks, VRAM release, and dictation UX shimmyshimmer 2026-07-17 07:39:15 -07:00
  • 1c7bce427e Revert "Feat/model picker per model config (#6647)" oobabooga 2026-07-17 07:38:46 -07:00
  • 709077c83c Fast-path the chat-template size check on character count danielhanchen 2026-07-17 13:44:07 +00:00
  • 4895a7c67a [pre-commit.ci] auto fixes from pre-commit.com hooks pre-commit-ci[bot] 2026-07-17 13:39:48 +00:00
  • bdc1629fea Bound the protected key set during legacy config migration danielhanchen 2026-07-17 13:38:07 +00:00
  • 9259937b2e Cap chat_template_override size on the model load path danielhanchen 2026-07-17 13:38:07 +00:00
  • ca7e0cb527 Send the HF token as a header, not a URL query param danielhanchen 2026-07-17 13:38:07 +00:00
  • dda6a1db63 Surface HiDream I1 in the image model catalog Daniel Han 2026-07-17 13:08:53 +00:00
  • f78ad94857 Add the HiDream-I1 family to the image backend Daniel Han 2026-07-17 13:08:53 +00:00
  • 9174535616 Surface HiDream I1 in the image model catalog Daniel Han 2026-07-17 13:08:53 +00:00
  • e098ee52ea Add the HiDream-I1 family to the image backend Daniel Han 2026-07-17 13:08:53 +00:00
  • 8cbdfbe355
         Feat/model picker per model config (#6647) Eyera 2026-07-17 15:08:01 +02:00
  • ff888fd3da Surface HunyuanImage 2.1 in the image model catalog Daniel Han 2026-07-17 12:52:01 +00:00
  • a3992cd0b6 Add the HunyuanImage 2.1 family to the image backend Daniel Han 2026-07-17 12:52:01 +00:00
  • 5887d2489c Surface HunyuanImage 2.1 in the image model catalog Daniel Han 2026-07-17 12:52:01 +00:00
  • c2743eb697 Add the HunyuanImage 2.1 family to the image backend Daniel Han 2026-07-17 12:52:01 +00:00
  • 035cabd5c3 Wire the hosted Lumina Image 2.0 int8/fp8 checkpoints Daniel Han 2026-07-17 12:20:30 +00:00
  • 98673456c8 Wire the hosted Lumina Image 2.0 int8/fp8 checkpoints Daniel Han 2026-07-17 12:20:30 +00:00
  • 5c2bbf50a9 install: constrain kept torch pins to the supported window studio-cuda-torch-2110 Daniel Han 2026-07-17 11:38:35 +00:00
  • b13d5e3573 Surface Wan2.2-I2V-A14B in the video tab with a source-image control Daniel Han 2026-07-17 11:15:55 +00:00
  • cffee73135 Add Wan2.2-I2V-A14B image-to-video support to the video backend Daniel Han 2026-07-17 11:15:48 +00:00
  • 63cbf7878e Add the Lumina Image 2.0 family to the image catalog Daniel Han 2026-07-17 11:05:59 +00:00
  • e1896eeaad Resolve pre-quantized checkpoints per base variant Daniel Han 2026-07-17 11:07:17 +00:00
  • 26da79f390 install: keep the existing torch release when re-running the installer Daniel Han 2026-07-17 11:10:51 +00:00
  • 350e46bf2e Add the Lumina Image 2.0 family to the image catalog Daniel Han 2026-07-17 11:05:59 +00:00
  • 570cef6c79 Resolve pre-quantized checkpoints per base variant Daniel Han 2026-07-17 11:07:17 +00:00
  • 79daafedbd Merge remote-tracking branch 'origin/main' into r6959 Daniel Han 2026-07-17 11:01:01 +00:00
  • 5a58e9df13 Add FLUX.1 Krea dev to the image model catalog Daniel Han 2026-07-17 10:49:04 +00:00
  • 08b853df15 Add FLUX.1 Krea dev to the image model catalog Daniel Han 2026-07-17 10:49:04 +00:00
  • 44f100af97 Support LoRA adapters on torchao int8/fp8 quantized image pipelines Daniel Han 2026-07-17 09:37:39 +00:00
  • 362cacc448 Support LoRA adapters on torchao int8/fp8 quantized image pipelines Daniel Han 2026-07-17 09:37:17 +00:00
  • d5be6fb5a3 Add FLUX.2 Klein and FLUX.2-dev DiT LoRA training Daniel Han 2026-07-17 09:34:14 +00:00
  • 07076430b3 Add FLUX.2 Klein and FLUX.2-dev DiT LoRA training Daniel Han 2026-07-17 09:34:01 +00:00
  • b508c8fe89
         fix(save): unsloth_push_to_hub_gguf(save_method="lora") raises NameError (#7193) Andrew Chen 2026-07-17 16:55:34 +08:00
  • cf1910c92d Harden the diffusion memory plan against transient free-VRAM undercounts Daniel Han 2026-07-17 08:46:12 +00:00
  • b35985162f Harden the diffusion memory plan against transient free-VRAM undercounts Daniel Han 2026-07-17 08:46:12 +00:00
  • 06bd5a7e0b [pre-commit.ci] auto fixes from pre-commit.com hooks cuda-torch-index-override pre-commit-ci[bot] 2026-07-17 08:30:39 +00:00
  • a38580f2f5 install: compare major-only rocm pins, redact URL fragments, bound pinned CPU trio Daniel Han 2026-07-17 08:29:55 +00:00
  • 68ecf638f2 Make the selected model row clearly highlighted in the STT picker shimmyshimmer 2026-07-17 00:38:09 -07:00
  • a01c8b2529 Rename stale _starting references in the pump resilience tests shimmyshimmer 2026-07-17 00:16:34 -07:00
  • 6d33582b88 Keep Qwen-Image's text-stream linears bf16 on int8 (short prompts break torch._int_mm) Daniel Han 2026-07-17 07:16:27 +00:00
  • 5d8bf5d094 Keep Qwen-Image's text-stream linears bf16 on int8 (short prompts break torch._int_mm) Daniel Han 2026-07-17 07:16:23 +00:00
  • aaba2d0a22 install: re-apply a ROCm pin over an existing HIP wheel via the version tag Daniel Han 2026-07-17 07:00:51 +00:00
  • 162114e8b8 [pre-commit.ci] auto fixes from pre-commit.com hooks pre-commit-ci[bot] 2026-07-17 06:58:05 +00:00
  • e2c5c0e1a2 Merge remote-tracking branch 'origin/main' into studio-stt-dictation shimmyshimmer 2026-07-16 23:55:43 -07:00
  • b5e0404d9d Verify every shard of a sharded checkpoint in the downloaded check shimmyshimmer 2026-07-16 23:54:47 -07:00
  • ee0e059007 Route krea-2 through its per-component loader on the transformer-quant fast path Daniel Han 2026-07-17 06:36:27 +00:00
  • 91be795a9b Route krea-2 through its per-component loader on the transformer-quant fast path Daniel Han 2026-07-17 06:36:24 +00:00
  • b5d20b4035 Skip the duplicate source line for custom models in the STT picker shimmyshimmer 2026-07-16 23:06:29 -07:00
  • fbcf070fce Wire hosted pre-quantized DiT checkpoints into the image families Daniel Han 2026-07-17 05:47:45 +00:00
  • 5e1614259c Wire hosted pre-quantized DiT checkpoints into the image families Daniel Han 2026-07-17 05:47:45 +00:00
  • 7fb7e1675c Merge local engines into one option and source GGML models from unslothai shimmyshimmer 2026-07-16 22:32:40 -07:00
  • de0d5ce336 Add whisper.cpp dictation engine with per-engine downloads and history rework shimmyshimmer 2026-07-16 21:21:42 -07:00
  • 1777aae37e
         don't kill live llama-servers when a new Studio instance starts (#7182) Nilay 2026-07-17 04:54:05 +05:30
  • c4e6dd4f6c
         Studio: extract text from PDF web results (#7154) oobabooga 2026-07-16 19:48:12 -03:00
  • 3555dbdda7
         Studio: don't drop parallel tool calls after an internal no-op (#7157) oobabooga 2026-07-16 19:47:22 -03:00
  • 976b9466ed Cover nested canvas and dynamic import paths agent/harden-auto-permission-network-gates Michael Han 2026-07-16 15:06:05 -07:00
  • 7e6eacf322 Fix macOS permission browser helper agent/clarify-permission-mode-controls Michael Han 2026-07-16 14:48:53 -07:00
  • 29135c0ead Keep permission browser CI triggers reliable Michael Han 2026-07-16 14:37:09 -07:00
  • 124ada208f Ignore disabled RAG pill in compact count Michael Han 2026-07-16 14:34:48 -07:00
  • 13e9426e44 Harden sandbox import and canvas analysis Michael Han 2026-07-16 14:32:38 -07:00
  • 8c83478da0
         Studio: use one shared Hugging Face token across Settings and training (#7152) Nilay 2026-07-16 20:54:33 +05:30
  • 40c12f68c6 [pre-commit.ci] auto fixes from pre-commit.com hooks pre-commit-ci[bot] 2026-07-16 14:17:35 +00:00
  • 63967e1f2a
         Bump the actions group across 1 directory with 11 updates dependabot/github_actions/actions-d2ebdbca80 dependabot[bot] 2026-07-16 14:16:32 +00:00
  • 3122f832d9 install: split the survive-updates marker subsystem into a follow-up Daniel Han 2026-07-16 14:16:17 +00:00
  • 3d7b582fd6 Exercise permission controls across browsers Michael Han 2026-07-16 07:10:32 -07:00
  • 3be49070cc
         fix config cards clipping content at narrow window widths (#7146) Nilay 2026-07-16 19:40:30 +05:30
  • ec26768cbb Cover dynamic network imports and canvas edge cases Michael Han 2026-07-16 06:59:51 -07:00
  • 702f868f63
         Bump serde_with dependabot/cargo/studio/src-tauri/cargo-tauri-security-082059d156 dependabot[bot] 2026-07-16 13:45:18 +00:00
  • 14f17a7de8 Clarify tool permission controls Michael Han 2026-07-16 06:28:31 -07:00
  • 030f12753c
         Fix Inkling reasoning-effort coercion for duck-typed engine stand-ins (#7158) Daniel Han 2026-07-16 06:23:06 -07:00
  • e3674d6aa4
         CI: opt tool-calling smoke tests out of the chat tool approval gate (#7162) Daniel Han 2026-07-16 06:22:54 -07:00
  • eed25b3845 Harden auto permission network gates Michael Han 2026-07-16 06:22:16 -07:00
  • 18cf9c0381 Fix hidden cache row confirmation fix/hide-infra-models-hub-inventory Michael Han 2026-07-16 06:11:54 -07:00
  • aad11f4ef1
         Studio: clickable sidebar settings cog, long name truncation, Canvas menu opt-in (#7171) Michael Han 2026-07-16 06:01:37 -07:00
  • 26faceecf7
         Harden desktop release token permissions (#7172) Wasim Yousef Said 2026-07-16 14:01:48 +02:00
  • a024d84c53 [pre-commit.ci] auto fixes from pre-commit.com hooks pre-commit-ci[bot] 2026-07-16 11:49:49 +00:00