Commit Graph

  • b9e6016095 [pre-commit.ci] auto fixes from pre-commit.com hooks feature/docker-studio-v0.1.35-beta pre-commit-ci[bot] 2026-04-02 23:37:48 +00:00
  • 2cfb83dc71 fix: patch PEFT for Gemma4ClippableLinear in loader checkpoint path Roland Tannous 2026-04-02 23:24:47 +00:00
  • 44454091c0 add Docker support: skip venv, install only missing deps Roland Tannous 2026-03-27 19:39:37 +00:00
  • 6644a771b4 fix: patch PEFT for Gemma4ClippableLinear in loader checkpoint path (fixes export) (#4807) main Roland Tannous 2026-04-03 04:03:54 +04:00
  • f91ef8f9b0 fix(studio): lazy-import transformers in model_config to fix 5.x version switch (#4806) Roland Tannous 2026-04-03 02:56:01 +04:00
  • 737017a511 fix(studio): lazy-import extract_model_size_b in llama_cpp to fix transformers 5.x version switch fix/lazy-import-llama-cpp Roland Tannous 2026-04-02 21:42:56 +00:00
  • 8f721d28d6 Merge branch 'main' into pip pip Daniel Han 2026-04-02 12:28:40 -07:00
  • e553a8ad0b fix(studio): suppress fatal error when prebuilt manifest is missing (#4799) v0.1.35-beta Daniel Han 2026-04-02 12:18:11 -07:00
  • 90724f6dab fix(studio): suppress fatal error when prebuilt manifest is missing fix/llama-tag-latest Daniel Han 2026-04-02 19:14:00 +00:00
  • cd09070f40 build(deps): bump the pip group across 3 directories with 19 updates dependabot/pip/pip-06c0d1eadc dependabot[bot] 2026-04-02 19:13:13 +00:00
  • 396aa05ead Update pyproject.toml Daniel Han 2026-04-02 12:03:10 -07:00
  • 2297f73cad Merge branch 'main' into pip Daniel Han 2026-04-02 12:02:48 -07:00
  • 8ffd5826f2 Gemma-4 Daniel Han 2026-04-02 11:59:37 -07:00
  • 934478ae31 fix(studio): revert llama.cpp default tag to latest (#4797) Daniel Han 2026-04-02 11:52:37 -07:00
  • b6ffd94918 docs: add comment explaining latest vs master for llama.cpp tag Daniel Han 2026-04-02 18:51:30 +00:00
  • 1902e83a67 fix(studio): revert llama.cpp default tag to latest Daniel Han 2026-04-02 18:49:45 +00:00
  • 8ecfb07616 fix(studio): revert llama.cpp default tag to latest fix/llama-cpp-b8637 Daniel Han 2026-04-02 18:40:45 +00:00
  • 401621618b fix(studio): don't set trust_remote_code for Gemma 4 training (#4795) Daniel Han 2026-04-02 11:44:26 -07:00
  • 8d1712b4ea fix(studio): pin llama.cpp to b8637 release (Gemma 4 support) (#4796) Daniel Han 2026-04-02 11:43:53 -07:00
  • 7ae9b7f45f fix windows llama.cpp compile from source issue (#4793) DoubleMathew 2026-04-02 13:43:46 -05:00
  • 6d20440341 fix: address review feedback on install paths fix/studio-install-from-git-main Daniel Han 2026-04-02 18:12:05 +00:00
  • a484032f92 fix(studio): temporarily install unsloth and unsloth_zoo from git main Daniel Han 2026-04-02 18:02:59 +00:00
  • 7023e2a4ff fix(studio): prioritize curated defaults over HF download ranking in Recommended (#4792) Daniel Han 2026-04-02 10:46:53 -07:00
  • 0446d46689 fixed name (#4791) Roland Tannous 2026-04-02 21:04:42 +04:00
  • 1ce83c40aa fix(studio): build llama.cpp from master instead of latest release tag (#4790) Daniel Han 2026-04-02 09:45:56 -07:00
  • 2af53bf9a6 Pin transformers and huggingface-hub in main Studio venv (#4788) Daniel Han 2026-04-02 09:21:30 -07:00
  • a241c58d84 Use transformers v5.5-release branch and pin to 5.5.0 (#4786) Daniel Han 2026-04-02 09:10:02 -07:00
  • a353557249 Force llama.cpp to always use mainline ggml-org (#4785) Daniel Han 2026-04-02 09:03:00 -07:00
  • afb929271c fix: pass --clear to uv venv to handle leftover venv directory fix/uv-venv-clean Roland Tannous 2026-04-02 15:49:39 +00:00
  • 4fadb05c50 fix: detect Apple Silicon under Rosetta and use arm64 binaries fix/mac-llamacpp-binary Roland Tannous 2026-04-02 15:49:18 +00:00
  • f1c3b9caa9 Pin Gemma-4 transformers requirement to 5.5.0 stable (#4784) Daniel Han 2026-04-02 08:59:21 -07:00
  • 4f9986ecb9 fix(studio): improve tool-calling re-prompt for small models (#4783) Daniel Han 2026-04-02 08:59:02 -07:00
  • f9c4b08726 UI Changes (#4782) Daniel Han 2026-04-02 08:05:55 -07:00
  • 3b613eb1e8 ui improvement (#4781) Roland Tannous 2026-04-02 18:57:47 +04:00
  • f366d00451 some fixes fix/ui-fix Roland Tannous 2026-04-02 18:39:45 +04:00
  • c8d311a053 feat(studio): display images from Python tool execution in chat UI (#4778) Daniel Han 2026-04-02 05:08:16 -07:00
  • 9b5c98b89e studio: fix external models selectable in chat-only mode (#4779) feature/chat-api Lee Jackson 2026-04-02 13:06:26 +01:00
  • 5a5f1a4f34 studio: fix chat font changes leaking outside chat page (#4775) Lee Jackson 2026-04-02 13:04:23 +01:00
  • 66e84a33b8 unsloth gemma4 support files Roland Tannous 2026-04-02 15:54:59 +04:00
  • 1ce8a8e7cd Feat/custom llama prebuilt (#4771) DoubleMathew 2026-04-02 06:52:26 -05:00
  • b20efc370a Add regression tests for custom llama prebuilt installer (#4772) Daniel Han 2026-04-02 04:45:09 -07:00
  • e2fd946fe1 Add files via upload Michael Han 2026-04-02 03:00:10 -07:00
  • 31d6aeb197 Unsloth new logo Michael Han 2026-04-02 02:58:21 -07:00
  • 4cce8bc167 build(deps): bump the bun-frontend group across 1 directory with 16 updates dependabot/bun/studio/frontend/bun-frontend-87f39202d2 dependabot[bot] 2026-04-02 09:58:03 +00:00
  • bc283d2f1c build(deps): bump oxc-parser dependabot/npm_and_yarn/studio/backend/core/data_recipe/oxc-validator/npm-oxc-validator-da2c725985 dependabot[bot] 2026-04-02 09:56:35 +00:00
  • e4d1499230 fix(studio): prevent small models from stalling on tool-calling tasks (#4769) Daniel Han 2026-04-02 02:11:07 -07:00
  • dc0729aadf Add regression test for shell injection fix in GGML conversion (#4773) Daniel Han 2026-04-02 00:10:47 -07:00
  • 752cef3299 fix(security): shell injection in GGML export conversion (#4768) mateeaaaaaaa 2026-04-02 10:10:43 +03:00
  • ba8081fc96 fix(chat): correct loading text for cached models during inference (#4764) AdamPlatin123 2026-04-02 11:24:48 +08:00
  • ca4ea8b9fb studio: align composer/code, unify fonts, and remove tool collapse jitter (#4763) Lee Jackson 2026-04-02 03:57:10 +01:00
  • 19c2dd2fa4 Merge branch 'main' into feature/chat-api Roland Tannous 2026-04-01 23:52:51 +04:00
  • f8b789f063 feat(studio): preserve image parts in external provider chat requests Roland Tannous 2026-04-01 19:46:24 +00:00
  • 3cddb8f953 fix: persist session password in sessionStorage to survive page refreshes Roland Tannous 2026-04-01 19:19:18 +00:00
  • 1d882711da fix: cast PBKDF2 salt to BufferSource for strict TypeScript lib types Roland Tannous 2026-04-01 19:07:55 +00:00
  • 71b934ef9d Fix custom llama.cpp source builds and macos metal source builds (#4762) DoubleMathew 2026-04-01 14:06:39 -05:00
  • 593bd10506 feat(studio): encrypt external provider API keys at rest in localStorage Roland Tannous 2026-04-01 18:54:25 +00:00
  • 5508c16208 Merge branch 'main' into feature/chat-api Roland Tannous 2026-04-01 18:28:44 +00:00
  • 15cf301a5f fix: address PR review comments — system prompt visibility, retry loop, test logging Roland Tannous 2026-04-01 18:19:09 +00:00
  • 651a1f4ab5 add missing providers-api.ts file by Imagineer99 Roland Tannous 2026-04-01 18:03:09 +00:00
  • 39fe23ded8 Tests for architecture-aware KV cache estimation (#4760) Daniel Han 2026-04-01 06:13:37 -07:00
  • 653eb3819a fix(studio): allow context length slider to reach model's native limit (#4746) Daniel Han 2026-04-01 06:12:52 -07:00
  • d22b2a18f9 fix: add tokenizers to no-torch deps and TORCH_CONSTRAINT for arm64 macOS py313+ (#4748) Daniel Han 2026-04-01 06:12:17 -07:00
  • 76cb48be0b fix: studio web search SSL failures and empty page content (#4754) Daniel Han 2026-04-01 06:12:02 -07:00
  • f84c2d03d3 Add installer test coverage for prebuilt llama.cpp changes (#4756) Daniel Han 2026-04-01 06:06:29 -07:00
  • 428efc7d95 Resolve latest usable published llama.cpp release instead of fixed pinned tag (#4741) DoubleMathew 2026-04-01 08:06:17 -05:00
  • 5d7d882ce6 Fix save_pretrained_merged for full-finetuned models (#4755) Daniel Han 2026-04-01 06:05:37 -07:00
  • 77e1a9edc9 feat(studio): architecture-aware KV cache VRAM estimation (#4757) Daniel Han 2026-04-01 06:04:12 -07:00
  • 3f3757b143 Fix forward compatibility with transformers 5.x (#4752) Daniel Han 2026-04-01 06:04:03 -07:00
  • a9d6c26bb5 [pre-commit.ci] auto fixes from pre-commit.com hooks test/kv-cache-estimation-tests pre-commit-ci[bot] 2026-04-01 12:52:54 +00:00
  • f4efbba2a0 test: add MLA n_kv default safety test (66 tests total) Daniel Han 2026-04-01 12:51:55 +00:00
  • 87e5385b44 Fix MLA n_kv safety and use ceiling division for hybrid path fix/kv-cache-5path-estimation Daniel Han 2026-04-01 12:51:43 +00:00
  • 434dee9618 Tighten _can_estimate_kv gate and treat sliding_window=0 as disabled Daniel Han 2026-04-01 12:49:41 +00:00
  • 2fc61ebab3 Merge branch 'main' into fix/web-search-ssl-sni fix/web-search-ssl-sni Roland Tannous 2026-04-01 16:48:40 +04:00
  • ed836f143f fix: address PR review — use _create_connection override, reuse SSL context Roland Tannous 2026-04-01 12:42:42 +00:00
  • ae6fb93b6f Fix MLA fallback and SWA global/local ratio heuristic Daniel Han 2026-04-01 12:42:44 +00:00
  • 3182fd953c [pre-commit.ci] auto fixes from pre-commit.com hooks pre-commit-ci[bot] 2026-04-01 12:37:05 +00:00
  • 41198342d9 [pre-commit.ci] auto fixes from pre-commit.com hooks pre-commit-ci[bot] 2026-04-01 12:29:57 +00:00
  • bac2cfac7c feat(studio): architecture-aware KV cache VRAM estimation Daniel Han 2026-04-01 12:29:20 +00:00
  • bf55e0cba4 fix(studio): resolve SSL handshake failure in web_search tool calls Roland Tannous 2026-04-01 12:22:28 +00:00
  • 41df4ec437 feat(studio): strip org prefix in model search to surface unsloth variants (#4749) Roland Tannous 2026-04-01 15:37:28 +04:00
  • 391130f365 Merge remote-tracking branch 'origin/main' into feature/docker-studio-v0.1.3-beta feature/docker-studio-v0.1.3-beta Roland Tannous 2026-04-01 10:19:23 +00:00
  • c45d5b6efa Fix compute_loss token_type_ids propagation and ModernBERT flex_attention fix/transformers-v5-compat Daniel Han 2026-04-01 09:13:01 +00:00
  • 63ad6dbd6d Fix OOM model styling in Studio model selectors (#4738) Leo Borcherding 2026-04-01 04:06:49 -05:00
  • 6c0826a9e4 Fix Windows local GGUF model loading crash (#4730) Daniel Han 2026-04-01 01:38:09 -07:00
  • d32ee870a7 [pre-commit.ci] auto fixes from pre-commit.com hooks pre-commit-ci[bot] 2026-04-01 08:32:56 +00:00
  • 2290407d99 Pass token_type_ids and mm_token_type_ids through GRPO VLM path Daniel Han 2026-04-01 08:31:41 +00:00
  • 256c6e4884 Refactor flex attn to prefer flash if possible (#4734) Datta Nimmaturi 2026-04-01 13:00:21 +05:30
  • b506fbd86f [pre-commit.ci] auto fixes from pre-commit.com hooks pre-commit-ci[bot] 2026-04-01 07:28:55 +00:00
  • d63cc57e1e fix: clear tool status badge immediately after tool execution (#4733) Wasim Yousef Said 2026-04-01 09:28:38 +02:00
  • c0cf5f257c Fix forward compatibility with transformers 5.x Daniel Han 2026-04-01 07:28:24 +00:00
  • c3ad85a995 studio: add API provider UI and integrate wiring (#4737) Lee Jackson 2026-04-01 08:08:29 +01:00
  • 81926463d9 Review PR 4741 changes pr-4741-5147102380432514182 google-labs-jules[bot] 2026-04-01 06:53:13 +00:00
  • d1e38589eb Attribute is_rdna() RDNA2/3/3.5/4 expansion to PR #4428 feature/rocm-support-v2 Daniel Han 2026-04-01 06:48:10 +00:00
  • 4fb9778988 feat: move folder management into model selector dropdown (#4731) Wasim Yousef Said 2026-04-01 08:15:50 +02:00
  • 2cac3e8e4d studio: Polish Windows installer/setup logs (#4736) Lee Jackson 2026-04-01 07:12:42 +01:00
  • d96e3a7096 fix(vram): use training-aware estimates and backend arch-based VRAM for model fitness fix/fitness-oom-calculator Roland Tannous 2026-04-01 04:55:51 +00:00
  • 405bb58284 Fix config stamping when SDPA is not supported flexattn_refactor_fixed Daniel Han 2026-03-31 17:29:07 +00:00
  • 3b8860c039 Consistent config stamping in vision.py attention fallbacks Daniel Han 2026-03-31 17:17:37 +00:00
  • f77a42e0ab Fix attention selection regressions in PR #4250 Daniel Han 2026-03-31 16:47:48 +00:00
  • b0b6c4daf9 Trim whitespace in JSONL export content feat/chat-ui-enhancements Daniel Han 2026-03-31 14:24:52 +00:00