Commit Graph

  • ff23dd343f
    mlx: apply repeat penalties in sampler (#15631) main Daniel Hiltgen 2026-04-18 07:49:38 -07:00
  • 123b300af6
    docs: update hermes (#15655) Parth Sareen 2026-04-17 14:20:59 -07:00
  • 57653b8e42
    cmd/launch: show WSL guidance on Windows instead of handing off (#15637) Parth Sareen 2026-04-16 17:18:04 -07:00
  • a50ce61c54
    launch: skip unchanged managed-single rewrite (#15633) Parth Sareen 2026-04-16 16:20:42 -07:00
  • 2bb7ea00d2
    create: avoid gc race with create (#15628) Daniel Hiltgen 2026-04-16 13:29:16 -07:00
  • 55fa80d07a
    mlx: additional gemma4 cache fixes (#15607) Daniel Hiltgen 2026-04-16 13:07:19 -07:00
  • b9cb535407
    mlx: fix gemma4 cache to use logical view (#15617) Daniel Hiltgen 2026-04-16 11:54:30 -07:00
  • 031baef094
    mlx: fix imagegen lookup (#15588) Daniel Hiltgen 2026-04-16 10:39:00 -07:00
  • 7d271e6dc9
    cmd/launch: add Copilot CLI integration (#15583) Mike Wallio 2026-04-15 20:22:53 -04:00
  • c88dae2d6b
    Merge pull request #15612 from ollama/drifkin/gemma4-split-templates Devon Rifkin 2026-04-15 17:15:35 -07:00
  • 9e3618d663 make empty block conditional Devon Rifkin 2026-04-15 15:35:25 -07:00
  • 5d920cc6bc
    Keep Gemma4 router projection in source precision (#15613) Daniel Hiltgen 2026-04-15 15:04:23 -07:00
  • e585ecd11f gemma4: render differently based on model size Devon Rifkin 2026-04-15 14:37:16 -07:00
  • cdddea0592
    launch: always list cloud recommendations first (#15593) Eva H 2026-04-15 13:17:35 -07:00
  • 43f90def04
    launch: add hermes (#15569) Parth Sareen 2026-04-15 12:00:23 -07:00
  • 06ae6367bd
    mlx: fix RotatingKVCache.concat() dropping context on mid-rotation (#15591) Daniel Hiltgen 2026-04-14 18:29:06 -07:00
  • 48ad7085c4
    mlx: Improve gemma4 performance with fused operations (#15587) Daniel Hiltgen 2026-04-14 18:04:04 -07:00
  • e1e3cec8d0 models: fuse MLP activation functions via mlx_compile Jesse Gross 2026-04-13 12:20:39 -07:00
  • d3e67e305c mlx: add compiled closure support Jesse Gross 2026-04-13 12:20:33 -07:00
  • 698e04a14b
    launch: OpenCode inline config (#15586) Eva H 2026-04-14 15:08:42 -07:00
  • 1d9537bc33
    launch/openclaw: fix --yes flag behaviour to skip channels configuration (#15589) Eva H 2026-04-14 13:57:35 -07:00
  • 120424d832
    Revert "launch/opencode: use inline config (#15462)" (#15568) Eva H 2026-04-13 18:40:17 -07:00
  • 5818001610
    launch: skip unchanged integration rewrite configration (#15491) Eva H 2026-04-13 17:18:56 -07:00
  • 2cba7756c5
    Gemma4 on MLX (#15244) Daniel Hiltgen 2026-04-13 16:36:51 -07:00
  • bf2a421727
    gemma4: restore e2b-style nothink prompt (#15560) Devon Rifkin 2026-04-13 14:26:15 -07:00
  • f3cf6b75fb
    launch/opencode: use inline config (#15462) Eva H 2026-04-13 13:41:31 -07:00
  • 5dfac387a6
    Revert "gemma4: fix nothink case renderer (#15553)" (#15556) Devon Rifkin 2026-04-13 13:12:18 -07:00
  • a99e5d9c22
    mac: prevent generate on cross-compiles (#15120) Daniel Hiltgen 2026-04-13 13:04:58 -07:00
  • 0abf3aca36
    cgo: suppress deprecated warning to quiet down go build (#15438) Daniel Hiltgen 2026-04-13 13:04:11 -07:00
  • ee0266462a
    Revert "gemma4: add nothink renderer tests (#15554)" (#15555) Devon Rifkin 2026-04-13 13:00:59 -07:00
  • c88fb286ec
    mlx: add op wrappers for Conv2d, Pad, activations, trig, and masked SDPA (#14913) Daniel Hiltgen 2026-04-13 11:43:24 -07:00
  • d3da29cbfc
    mlx: mixed-precision quant and capability detection improvements (#15409) Daniel Hiltgen 2026-04-13 11:43:07 -07:00
  • 1b70bb8a10
    gemma4: add nothink renderer tests (#15554) Devon Rifkin 2026-04-13 11:38:19 -07:00
  • ec29ce4ce3
    gemma4: fix compiler error on metal (#15550) Daniel Hiltgen 2026-04-13 11:32:00 -07:00
  • 4d75f5da03
    gemma4: fix nothink case renderer (#15553) Devon Rifkin 2026-04-13 11:23:19 -07:00
  • 798fd09bfe
    Update to ROCm 7.2.1 (#15483) saman-amd 2026-04-12 15:11:58 -04:00
  • 9330bb9120
    gemma4: be less strict about whitespace before bare keys (#15494) Devon Rifkin 2026-04-11 16:30:27 -07:00
  • 40a1317dfd
    gemma4: update renderer to match new jinja template (#15490) Devon Rifkin 2026-04-10 15:45:27 -07:00
  • fdfe9cec98
    model/parsers: fix missing parallel tool call indices (#15467) Devon Rifkin 2026-04-10 15:23:21 -07:00
  • 9517864603
    app/ui: re-validate image attachments when selected model changes (#15272) Matteo Celani 2026-04-10 23:03:51 +02:00
  • 8e6d86dbe3
    docs: add hermes agent integration guide (#15488) Bruce MacDonald 2026-04-10 13:13:36 -07:00
  • 80d3744c5d
    launch: update openclaw channel message (#15463) Parth Sareen 2026-04-09 15:20:30 -07:00
  • 2a94f03823
    launch: add re-run hint to dependency error message (#15439) Eva H 2026-04-09 09:51:34 -07:00
  • eb97274e5c
    modelfiles: fix /save command and add shortname for safetensors based models (#15413) Patrick Devine 2026-04-08 21:05:39 -07:00
  • 6b5db12aa2
    mlx: remove stale x86 libmlx library (#15443) Daniel Hiltgen 2026-04-08 20:51:47 -07:00
  • 612f0a17d3
    fix: improve error message for unknown input item type in responses API (#15424) 7. Sun 2026-04-09 01:41:12 +01:00
  • 673726fa0e
    app: restore launch default and refine launch sidebar open for app (#15437) Parth Sareen 2026-04-08 16:59:21 -07:00
  • b5918f9785
    pull/push: refine safetensors (#14946) Daniel Hiltgen 2026-04-08 14:15:39 -07:00
  • d17f482d50
    launch/opencode: detect curl installed opencode at ~/.opencode/bin (#15197) Eva H 2026-04-08 13:54:51 -07:00
  • 4e16f562c0
    launch: add openclaw channels setup (#15407) Parth Sareen 2026-04-08 13:25:27 -07:00
  • 55308f1421
    launch: update ctx length for glm-5.1 and gemma4 (#15411) Parth Sareen 2026-04-08 12:11:50 -07:00
  • d64812eb5d
    cmd: improve multi-select sorting and selection status (#15200) Eva H 2026-04-08 10:39:18 -07:00
  • f86a969f27
    responses: add support for fn call output arrays (#15406) Devon Rifkin 2026-04-07 16:47:30 -07:00
  • 9fa80a1660
    app/ui: fix lint errors for unused vars, prefer-const, and empty catch (#15282) Matteo Celani 2026-04-08 01:28:36 +02:00
  • dde09129d1
    gemma4: Disable FA on older GPUs where it doesn't work (#15403) Daniel Hiltgen 2026-04-07 14:54:25 -07:00
  • 780556c4d0
    mlx: use default http client (#15405) Patrick Devine 2026-04-07 14:53:23 -07:00
  • dfae363b5b
    gemma4: add missing file (#15394) Daniel Hiltgen 2026-04-07 09:18:01 -07:00
  • 30fdd229a4
    create: Clean up experimental paths, fix create from existing safetensor model (#14679) Daniel Hiltgen 2026-04-07 08:12:57 -07:00
  • e823bff873
    gemma4: enable flash attention (#15378) Daniel Hiltgen 2026-04-07 08:12:36 -07:00
  • 8968740836
    mlx: Improve M5 performance with NAX (#15345) Daniel Hiltgen 2026-04-07 08:12:24 -07:00
  • 8c8f8f3450
    model/parsers: add gemma4 tool call repair (#15374) Devon Rifkin 2026-04-06 18:47:17 -07:00
  • 82f0139587
    launch/openclaw: patch approvedScopes baseline for TUI pairing (#15375) Parth Sareen 2026-04-06 18:00:12 -07:00
  • 26a58b294c
    app: update featured models (#15373) Bruce MacDonald 2026-04-06 16:35:35 -07:00
  • 34a790a2e6
    model/parsers: suppress extra gemma4 closing tool tags (#15370) Devon Rifkin 2026-04-06 12:41:33 -07:00
  • 4589fa2cf5
    app: default app home view to new chat instead of launch (#15312) Jeffrey Morgan 2026-04-03 21:50:55 -07:00
  • 4bc2728047
    Revert "enable flash attention for gemma4 (#15296)" (#15311) Daniel Hiltgen 2026-04-03 17:44:44 -07:00
  • 49d5fd5a3e
    model/parsers: rework gemma4 tool call handling (#15306) Devon Rifkin 2026-04-03 14:35:00 -07:00
  • 3cd2b03a5e ggml: fix ROCm build for cublasGemmBatchedEx reserve wrapper Jesse Gross 2026-04-03 13:26:50 -07:00
  • c8e0878814
    enable flash attention for gemma4 (#15296) Daniel Hiltgen 2026-04-03 12:46:18 -07:00
  • bb0c58e134 ggml: skip cublasGemmBatchedEx during graph reservation Jesse Gross 2026-04-03 11:26:03 -07:00
  • 036ed1b9b5
    model/parsers: fix gemma4 arg parsing when quoted strings contain " (#15254) Devon Rifkin 2026-04-02 22:52:51 -07:00
  • 3536ef58f6
    bench: add prompt calibration, context size flag, and NumCtx reporting (#15158) Daniel Hiltgen 2026-04-02 14:23:53 -07:00
  • de9673ac3f
    tokenizer: add byte fallback for SentencePiece BPE encoding (#15232) Daniel Hiltgen 2026-04-02 13:04:45 -07:00
  • 96b202d34b
    Add support for gemma4 (#15214) Daniel Hiltgen 2026-04-02 11:33:33 -07:00
  • 79865e6c5a
    app: use the same client for inference and other requests (#15204) Devon Rifkin 2026-04-02 11:07:50 -07:00
  • 5ab10d347a
    app: add launch page for a simple way to launch integrations (#15182) Parth Sareen 2026-04-02 10:31:19 -07:00
  • a8292dd85f
    launch: replace deprecated OPENAI_BASE_URL with config.toml profile for codex (#15041) Eva H 2026-04-01 08:43:23 -07:00
  • cb0033598e
    tokenizer: add SentencePiece-style BPE support (#15162) Daniel Hiltgen 2026-03-31 17:00:36 -07:00
  • 4d14b0ff92
    mlx: respect tokenizer add_bos_token setting in pipeline (#15185) Daniel Hiltgen 2026-03-31 16:46:30 -07:00
  • d9cb70c270
    docs: update pi docs (#15152) Parth Sareen 2026-03-31 16:37:55 -07:00
  • 31f968fe1f
    cmd: set OpenCode default model in config (#15127) Jeffrey Morgan 2026-03-29 12:11:36 -07:00
  • b7bda92d52
    model: add qwen3-next compatibility for legacy ssm_in projections (#15133) Jeffrey Morgan 2026-03-29 11:50:47 -07:00
  • 8e54823fd3
    revert context length warnings change (#15121) Parth Sareen 2026-03-28 16:43:59 -07:00
  • 7c8da5679e
    launch: improve multi-select for already added models (#15113) Parth Sareen 2026-03-28 13:44:40 -07:00
  • 6214103e66
    launch: auto-install pi and manage web-search lifecycle (#15118) Parth Sareen 2026-03-28 13:06:20 -07:00
  • 9e7cb9697e
    mlx: fix vision capability + min version (#15106) Patrick Devine 2026-03-27 17:09:28 -07:00
  • 3824e380a8
    server: preserve raw manifest bytes during pull (#15104) Bruce MacDonald 2026-03-27 15:42:31 -07:00
  • c9b2dcfc52
    anthropic: fix empty inputs in content blocks (#15105) Devon Rifkin 2026-03-27 15:41:27 -07:00
  • b00bd1dfd4
    launch: skip context length warning for MLX models and show model name (#15102) Parth Sareen 2026-03-27 15:01:33 -07:00
  • ac83ac20c4 anthropic: fix KV cache reuse degraded by tool call argument reordering Jesse Gross 2026-03-09 16:24:57 -07:00
  • e7ccc129ea
    app: fix false "out of date" model warnings (#15101) Bruce MacDonald 2026-03-27 14:15:10 -07:00
  • 69ed0c2729
    parsers: qwen3.5 streaming tool-call parsing and add regression test (#15098) Jeffrey Morgan 2026-03-27 14:04:14 -07:00
  • 1cefa749aa
    model/parsers: close think block if tool block starts in Qwen3.5 (#15022) Alfredo Matas 2026-03-27 19:28:34 +01:00
  • aec2fef95d
    ci: harden cuda include path handling (#15093) Daniel Hiltgen 2026-03-27 07:57:07 -07:00
  • 366625a831
    launch: warn when server context length is below 64k for local models (#15044) Eva H 2026-03-27 00:15:53 -07:00
  • 516ebd8548
    ci: include mlx jit headers on linux (#15083) Daniel Hiltgen 2026-03-26 23:10:07 -07:00
  • f567abc63f
    tui: update chat title (#15082) Parth Sareen 2026-03-26 18:06:53 -07:00
  • 1adfc27f04
    launch/vscode: prefer known vs code paths over code on PATH (#15073) Eva H 2026-03-26 15:06:28 -07:00
  • 4a2b9f9dbc
    launch: hide cline integration (#15080) Parth Sareen 2026-03-26 14:33:43 -07:00
  • e46b67a6cc
    launch: hide vs code (#15076) Parth Sareen 2026-03-26 13:52:50 -07:00