Commit Graph

  • 7563f51fdf sample: fix lint issues in penalty sampler jmorganca/faster-reptition-penalties jmorganca 2026-03-10 22:08:39 -07:00
  • 4ad9f16d68 Fix repeat penalty default jmorganca 2026-03-10 21:45:29 -07:00
  • d64d0d235d Add sampler reset on reprocess jmorganca 2026-03-10 21:37:09 -07:00
  • 54e05172a0 Revert "runner: add token history sampling parameters to ollama runner (#14537)" (#14776) Jeffrey Morgan 2026-03-10 21:07:52 -07:00
  • 4fdeb59325 convert: handle layers_block_type config field jmorganca/convert-nemotron jmorganca 2026-03-10 20:51:52 -07:00
  • 464186e995 config: qwen3.5 recommendations (#14758) Parth Sareen 2026-03-10 18:04:57 -07:00
  • 8c4d5d6c2f cloud_proxy: send ollama client version (#14769) Devon Rifkin 2026-03-10 15:53:25 -07:00
  • bc72b14016 docs: update claude code docs (#14770) Parth Sareen 2026-03-10 15:52:41 -07:00
  • 61086083eb server: add experimental web search and web fetch routes (#14753) Parth Sareen 2026-03-09 21:52:12 -07:00
  • 62d1f01ab4 ci: Fix windows build (#14754) v0.17.8-rc1 Daniel Hiltgen 2026-03-09 19:27:59 -07:00
  • 10e51c5177 MLX: add header vendoring and remove go build tag (#14642) v0.17.8-rc0 Daniel Hiltgen 2026-03-09 17:24:45 -07:00
  • 3e06bde643 mlx: get parameters from modelfile during model creation (#14747) Patrick Devine 2026-03-09 15:33:24 -07:00
  • 6be2de8214 app: auto update should be enabled when reset to defaults (#14741) Eva H 2026-03-09 15:02:36 -04:00
  • ebb1b9ec14 rocm: update linux to v7.2 (#14391) Daniel Hiltgen 2026-03-09 08:26:55 -07:00
  • d126467d5d x/mlxrunner: replace sampler interface chain with single stateful Sampler (#14652) Patrick Devine 2026-03-07 17:50:57 -08:00
  • afb4c62fbf cloud_proxy: handle stream disconnects gracefully (#14685) Devon Rifkin 2026-03-06 19:18:52 -08:00
  • e790dc435b mlx: int4 groupsize 64 (#14682) Patrick Devine 2026-03-06 16:39:47 -08:00
  • 288077c3a3 build: smarter docker parallelism (#14653) Daniel Hiltgen 2026-03-06 16:36:22 -08:00
  • 4425c54eda create: fix localhost handling (#14681) Daniel Hiltgen 2026-03-06 16:35:58 -08:00
  • 778899a5d2 docs: format compat docs (#14678) Michael Yang 2026-03-06 14:53:17 -08:00
  • 4eab60c1e2 Reapply "don't require pulling stubs for cloud models" again (#14608) Jeffrey Morgan 2026-03-06 14:27:47 -08:00
  • 1af850e6e3 parsers: repair unclosed arg_value tags in GLM tool calls (#14656) Bruce MacDonald 2026-03-06 14:08:34 -08:00
  • 9b0c7cc7b9 cmd: override stale entries for context window pi (#14655) v0.17.7-rc2 v0.17.7 Parth Sareen 2026-03-05 16:30:24 -08:00
  • 6928630601 mlx: prevent remote creation mismatch (#14651) Daniel Hiltgen 2026-03-05 14:59:00 -08:00
  • 9896e3627f cmd/config: fix cloud model limit lookups in integrations (#14650) v0.17.7-rc1 Parth Sareen 2026-03-05 13:57:28 -08:00
  • 15732f0ea7 cmd: use native Ollama API endpoint for OpenClaw (#14649) Bruce MacDonald 2026-03-05 13:29:17 -08:00
  • 562c76d7cc cmd: add qwen3.5 context length for launch (#14626) v0.17.7-rc0 Parth Sareen 2026-03-04 14:10:52 -08:00
  • 122c68c151 server: loosen thinking level constraint (#14625) Parth Sareen 2026-03-04 13:42:18 -08:00
  • 947200698f cmd/pi: fix happy path install parth-pi-thinking ParthSareen 2026-03-04 03:00:31 -08:00
  • 82848a7806 model: fix renderer and parser for qwen3.5 (#14605) v0.17.6 Jeffrey Morgan 2026-03-03 20:58:29 -08:00
  • 39982a954e Revert "Reapply "don't require pulling stubs for cloud models"" (#14606) Jeffrey Morgan 2026-03-03 20:56:10 -08:00
  • e9f6ea232f Add qwen3.5-next-moe support to MLX runner and models (#14417) Patrick Devine 2026-03-03 16:39:22 -08:00
  • 110eff01a9 chore: remove old imagegen LLMs models (#14597) Patrick Devine 2026-03-03 13:23:40 -08:00
  • 799e51d419 Reapply "don't require pulling stubs for cloud models" Jeffrey Morgan 2026-03-03 13:17:10 -08:00
  • e8fcb29586 model/renderers: fix glm-ocr image tags in renderer prompts (#14584) Victor-Quqi 2026-03-04 04:51:34 +08:00
  • 97d2f05a6d Revert "don't require pulling stubs for cloud models (#14574)" (#14596) Jeffrey Morgan 2026-03-03 12:51:23 -08:00
  • 8207e55ec7 don't require pulling stubs for cloud models (#14574) Devon Rifkin 2026-03-03 10:46:33 -08:00
  • ad16bffc7d mlx: Remove peak memory from the API Jesse Gross 2026-03-02 11:37:52 -08:00
  • c1e3ef4bcc mlxrunner: Refcount pinned tensors Jesse Gross 2026-03-02 12:48:02 -08:00
  • a3093cd5e5 cmd/opencode: rename provider from "Ollama (local)" to "Ollama" (#14566) Parth Sareen 2026-03-02 14:17:18 -08:00
  • 23d4cad1a2 server: verify digest is not empty on create (#14555) Bruce MacDonald 2026-03-02 13:43:35 -08:00
  • 86513cb697 runner: add token history sampling parameters to ollama runner (#14537) v0.17.5 Jeffrey Morgan 2026-03-01 19:16:07 -08:00
  • 3490e9590b model/qwen3next: avoid crash in in DeltaNet when offloading (#14541) Jeffrey Morgan 2026-03-01 18:44:04 -08:00
  • 67ce53b9b5 wip sampling pdevine/sampling-penalties Patrick Devine 2026-02-27 21:43:19 -08:00
  • dd497534c4 allow think/nothink in mlxrunner Patrick Devine 2026-02-26 23:29:04 -08:00
  • 560626fb43 cleanup Patrick Devine 2026-02-26 17:28:38 -08:00
  • 1a23c1a810 add qwen3.5 Patrick Devine 2026-02-25 16:21:42 -08:00
  • a6c1aa4da5 smaller recurrent cache Patrick Devine 2026-02-25 15:50:41 -08:00
  • 330b19b73f server: chunk quantization writes to reduce create memory usage jmorganca/fix-create-quantization-memory jmorganca 2026-02-28 23:21:37 -08:00
  • 8da09b1e7e qwen3next: add compatibility with imported GGUF models (#14517) Jeffrey Morgan 2026-02-28 14:21:42 -08:00
  • a60b9adcce mlxrunner: Fix prompt eval timing and count metrics Jesse Gross 2026-02-26 14:45:14 -08:00
  • a16f96658b mlxrunner: Enforce model context limit Jesse Gross 2026-02-25 15:07:09 -08:00
  • 18ab09b431 mlxrunner: Propagate pipeline errors to client via api.StatusError Jesse Gross 2026-02-26 12:23:06 -08:00
  • 638faeac54 mlxrunner: Report actual memory usage from runner Jesse Gross 2026-02-25 15:06:37 -08:00
  • 241ac6e003 check with HEAD before upload dongchen/resumable_transfer_fix Dong Chen 2026-02-27 14:29:17 -08:00
  • 17dc6918cf handle edge case when a blob has been uploaded earlier Dong Chen 2026-02-11 09:34:00 -08:00
  • dd5eb6337d mlxrunner: Fix panic on full KV cache hit Jesse Gross 2026-02-25 17:03:31 -08:00
  • 79917cf80b show peak memory usage (#14485) Patrick Devine 2026-02-26 18:38:27 -08:00
  • 857cffd22a bugfix: fix crash bug in token cache logic pdevine/sampling-cache-error Patrick Devine 2026-02-26 18:29:40 -08:00
  • cc90a035a0 model/parsers: add stable tool call indexing for glm47 and qwen3 parsers (#14484) v0.17.4 Parth Sareen 2026-02-26 18:14:29 -08:00
  • d98dda4676 model: fix qwen3 tool calling in thinking (#14477) v0.17.3 Jeffrey Morgan 2026-02-26 16:13:18 -08:00
  • d69ddc1edc fix: window app crash on startup when update is pending (#14451) v0.17.2 Eva H 2026-02-26 16:47:12 -05:00
  • 9bf41969f0 app: fix first update check delayed by 1 hour (#14427) v0.17.1 Eva H 2026-02-25 18:29:55 -05:00
  • 4d5ff25724 mlxrunner: Report actual memory usage from runner jessegross/mlx-usage Jesse Gross 2026-02-25 15:06:37 -08:00
  • 0f23b7bff5 mlxrunner: Cancel in-flight requests when the client disconnects Jesse Gross 2026-02-24 14:19:33 -08:00
  • 4e57d2094e mlxrunner: Simplify pipeline memory and cache management Jesse Gross 2026-02-24 14:19:12 -08:00
  • 7f9efd53df model: add support for qwen3.5-27b model (#14415) v0.17.1-rc2 Jeffrey Morgan 2026-02-25 01:09:58 -08:00
  • da70c3222e model: support for qwen3.5 architecture (#14378) v0.17.1-rc1 Jeffrey Morgan 2026-02-24 20:08:05 -08:00
  • 9d902d63ce ggml: ensure tensor size is valid (#14406) Bruce MacDonald 2026-02-24 21:52:44 -04:00
  • 99e470fe9f cmd: config update to use native Ollama API for OpenClaw hoyyeva/openclaw-config Eva Ho 2026-02-24 17:26:39 -05:00
  • d0c96be74e ui: santize HTML in chat markdown hoyyeva/app-html Eva Ho 2026-02-24 12:57:25 -05:00
  • f4f0a4a471 update mlx-c bindings to 0.5.0 (#14380) v0.17.1-rc0 Daniel Hiltgen 2026-02-23 16:44:29 -08:00
  • 3323c1d319 app: add upgrade configuration to settings page (#13512) Eva H 2026-02-23 18:08:52 -05:00
  • f20dc6b698 mlx: don't default to affine quantization for unquantized models Jesse Gross 2026-02-18 15:32:40 -08:00
  • 4b2ac1f369 model: improvements to LFM architectures (#14368) Jeffrey Morgan 2026-02-23 14:38:10 -08:00
  • 8daf47fb3a mlxrunner: Fix duplicate log prefixes and reduce log noise Jesse Gross 2026-02-23 13:28:01 -08:00
  • 6c980579cd ui: use capability-based detection for web search (#14336) Eva H 2026-02-23 15:00:09 -05:00
  • 5c73c4e2ee mlxrunner: Simplify KV cache to single-entry prefix matching Jesse Gross 2026-02-19 16:50:18 -08:00
  • 5daf59cc66 mlxrunner: Fix memory leaks with pin/sweep lifecycle management Jesse Gross 2026-02-19 15:05:35 -08:00
  • 0ade9205cc models: add nemotronh architecture support (#14356) Jeffrey Morgan 2026-02-22 15:09:14 -08:00
  • 06edabdde1 cmd/config: install web search plugin to user-level extensions dir (#14362) v0.17.0-rc2 v0.17.0 Parth Sareen 2026-02-22 02:17:03 -08:00
  • 8b4e5a82a8 mlx: remove noisy error output from dynamic library loading (#14346) v0.17.0-rc1 Jeffrey Morgan 2026-02-20 23:46:07 -08:00
  • 3445223311 cmd: openclaw onboarding (#14344) v0.17.0-rc0 Parth Sareen 2026-02-20 19:08:38 -08:00
  • fa6c0127e6 app: expose server's default context length to UI (#14037) Jeffrey Morgan 2026-02-20 18:56:30 -08:00
  • 00f67e807a Add qwen3.5-next-moe support to MLX runner and models pdevine/qwen3next Patrick Devine 2026-02-18 09:56:54 -08:00
  • 97323d1c68 consolidate the tokenizer (#14327) Patrick Devine 2026-02-19 15:55:45 -08:00
  • 458dd1b9d9 mlx: try loading library via rpath before searching directories (#14322) natl-set 2026-02-20 01:55:02 +07:00
  • 917b131636 Apply suggestion from @BruceMacD brucemacd/sign-sh-install Bruce MacDonald 2026-02-19 10:00:13 -08:00
  • 365a3657ad fix test home on windows brucemacd/tui-update Bruce MacDonald 2026-02-18 18:37:24 -08:00
  • 9d02d1d767 install: prevent partial download script execution (#14311) v0.16.3-rc2 v0.16.3 Bruce MacDonald 2026-02-18 18:32:45 -08:00
  • 71c1d8d0a9 cmd: ollama update Bruce MacDonald 2026-02-18 16:41:59 -08:00
  • 1a636fb47a cmd: set codex env vars on launch and handle zstd request bodies (#14122) Bruce MacDonald 2026-02-18 17:19:36 -08:00
  • 0759fface9 Revert "chore: update mlx-c bindings to 0.5.0 (#14303)" (#14316) Patrick Devine 2026-02-18 17:01:25 -08:00
  • 34515870ce install: check team identifier after download Bruce MacDonald 2026-02-18 13:23:14 -08:00
  • 325b72bc31 cmd/tui: default to single-select for editor integrations (#14302) v0.16.3-rc1 Parth Sareen 2026-02-17 18:17:27 -08:00
  • f01a9a7859 chore: update mlx-c bindings to 0.5.0 (#14303) Patrick Devine 2026-02-17 16:48:16 -08:00
  • 9ac1300805 fix lint brucemacd/usage-api Bruce MacDonald 2026-01-28 13:16:52 -08:00
  • 43d9907dd6 fix tests Bruce MacDonald 2026-01-28 13:07:48 -08:00
  • 91dc088e8b server: usage api Bruce MacDonald 2026-01-27 17:01:18 -08:00
  • 9aefd2dfee model: add qwen3 support to mlxrunner (#14293) v0.16.3-rc0 Patrick Devine 2026-02-17 13:58:49 -08:00