| pr |
0.75 |
Fix inaccurate num_ctx documentation description |
|
Overly formal, polite, and ChatGPT-like |
2026-04-14 |
| pr |
0.40 |
Add llm-box to Terminal & CLI section |
|
Some promotional phrases, but likely use |
2026-06-30 |
| pr |
0.30 |
ml/backend/ggml: add experimental SYCL backend support |
|
Draft language and meta-commentary, but |
2026-05-07 |
| pr |
0.20 |
Follow-up: metrics formatting and launch test lint cleanup |
|
Human merge strategy and domain-specific |
2026-07-04 |
| pr |
0.20 |
Follow-up: metrics formatting and launch test lint cleanup |
|
Human merge strategy and domain-specific |
2026-07-04 |
| pr |
0.20 |
fix: flush OpenAI-compatible streaming responsesfix: flush O |
|
Some formal language but very technical |
2026-07-04 |
| pr |
0.20 |
docs: add Dakera to Community Integrations |
|
Contains some detailed explanation, but |
2026-07-01 |
| pr |
0.20 |
fix(cloud): accept x-api-key header for Anthropic API compat |
|
Mostly changelog, with some template str |
2026-06-26 |
| pr |
0.20 |
feat llm: update llama.cpp HEAD to add powerpc as a supporte |
|
Slightly formal but with issue links and |
2026-06-19 |
| pr |
0.20 |
fix cuda toolkit lookup and parallel |
|
Polished, long-form rationale, but domai |
2026-06-07 |
| pr |
0.20 |
Fix default versions to be proper semver |
|
Technical, domain-specific, informal sen |
2026-06-30 |
| pr |
0.20 |
docs: add CorvinOS to Bots & Messaging integrations |
|
Concise style and domain terms; minimal |
2026-06-27 |
| commit |
0.15 |
mlx: x/create rewrite (#16919) |
|
Technical breakdown with domain specific |
2026-07-04 |
| pr |
0.15 |
mlx: x/create rewrite |
|
Informal, technical with step breakdown; |
2026-06-26 |
| commit |
0.10 |
MLX: wire up scheduler selected context size for ps (#16918) |
|
Technical detail and mild tone; no AI ha |
2026-06-26 |
| commit |
0.10 |
llm: decouple prompt caching from context shift (#16639) |
|
Direct, domain-specific language; no AI |
2026-06-11 |
| commit |
0.10 |
mlxrunner: harden linear/embedding layers against over-promo |
|
Concise technical phrasing and jargon; h |
2026-06-11 |
| commit |
0.10 |
mlx: "requires" in modelfile is being ignored for mlx based |
|
Technical, concise; just slightly formal |
2026-06-03 |
| commit |
0.10 |
log template details to aid troubleshooting (#16403) |
|
Concise technical explanation, natural f |
2026-06-01 |
| pr |
0.10 |
llm: clear parsed buffer sizes when resetting load accountin |
|
Technical language, specific, no AI hall |
2026-07-05 |
| pr |
0.10 |
create: fix parser/renderer mapping for qwen3 variants |
|
Specific bug description, domain-heavy, |
2026-04-02 |
| pr |
0.10 |
fix: include GGUF chat template tools capability in /api/tag |
|
Detailed, technical, informal, no AI phr |
2026-07-05 |
| pr |
0.10 |
server: return runner status code from EmbeddingsHandler |
|
Uses examples, tables, domain detail, na |
2026-07-05 |
| pr |
0.10 |
llm: sum sliding-window KV caches in VRAM accounting |
|
Specific bug/analysis, many domain detai |
2026-07-05 |
| pr |
0.10 |
server: add metrics endpoint |
|
Template sections filled with direct, mi |
2026-07-01 |
| pr |
0.10 |
fix: guard untrusted-input DoS panics in GGUF and registry p |
|
Direct, technical; informal notes and er |
2026-07-04 |
| pr |
0.10 |
llm: filter noisy llama-server logs |
|
Terse technical summary using project-sp |
2026-06-27 |
| pr |
0.10 |
XDG Spec Support |
|
Direct technical summary and changelist, |
2025-06-23 |
| pr |
0.10 |
Fix symlink path for ollama in install script |
|
Brief, domain-specific, and minor edit w |
2026-06-30 |
| pr |
0.10 |
llama: enable FA on CUDA CC 6.x GPUs |
|
Technical, concise, includes hardware/GP |
2026-07-01 |
| pr |
0.10 |
compat: use UTF-8-safe file open |
|
Concise technical description, specific |
2026-07-02 |
| pr |
0.10 |
llm: allow iGPU mmproj offload with fit padding |
|
Informal, dense technical explanation; d |
2026-07-01 |
| pr |
0.10 |
launch: add WebBrain handoff |
|
Technical, concise descriptions with cle |
2026-07-02 |
| pr |
0.10 |
launch: repair missing editor config before launch |
|
Domain-specific, informal, with some min |
2026-05-01 |
| pr |
0.10 |
Add new model servam 30b and 105b |
|
Sparse free-text, terse, domain content; |
2026-03-08 |
| pr |
0.10 |
fix(ui): re-encode clipboard/pasted images to PNG via canvas |
|
Technical, domain-specific, informal and |
2026-07-01 |
| pr |
0.10 |
server: grow embedding batch for long inputs |
|
Mainly technical, reviews in free-form a |
2026-06-24 |
| pr |
0.10 |
Fix macOS installer app launch path |
|
Detailed, technical, includes real CLI o |
2026-06-09 |
| pr |
0.10 |
cmake: ease Windows clang and custom ROCm builds from source |
|
Technical, domain-specific, no AI hallma |
2026-06-30 |
| pr |
0.10 |
fix: improve CLI image path recognition for escaped and tild |
|
Technical, concise summary of CLI parsin |
2026-06-19 |
| pr |
0.10 |
server: raise automatic context floor for vision models |
|
Bug summary, technical explanation, uses |
2026-06-30 |
| pr |
0.10 |
server: inherit model_info when creating from a remote model |
|
Technical explanation, no AI-typical phr |
2026-06-30 |
| pr |
0.10 |
feat(cache): write CACHEDIR.TAG into the models directory |
|
Detailed, technical explanation with dom |
2026-06-10 |
| pr |
0.10 |
fix(cmd): prevent ANSI escape sequences in piped output |
|
Natural technical discussion & context-s |
2026-06-22 |
| pr |
0.10 |
server: default mistral3 models to ministral parser |
|
Technical language and informal, terse c |
2026-06-26 |
| pr |
0.09 |
Add num_cpu_moe option to keep MoE expert tensors on the CPU |
|
Technical explanation with domain detail |
2026-06-12 |
| pr |
0.07 |
fs/ggml: fix KV cache over-estimation for sliding-window att |
|
Technical context, domain terms, table—n |
2026-06-29 |
| commit |
0.05 |
launch: auto-install opencode when missing (#16806) |
|
Terse, domain-specific commit message; h |
2026-06-19 |
| commit |
0.05 |
launch: auto-install Claude Code (#16802) |
|
Concise changelog phrasing; lacks AI sty |
2026-06-19 |
| commit |
0.05 |
server: context shift for context windows larger than 8k, ad |
|
Terse commit message, human style, no AI |
2026-06-15 |
| commit |
0.05 |
model: update lfm2 parser/renderer for optional thinking (#1 |
|
Brief technical wording, lacks AI-style |
2026-06-15 |
| commit |
0.05 |
llama: update llama.cpp to b9637 (#16609) |
|
Direct update message, domain-specific, |
2026-06-15 |
| pr |
0.05 |
server: notify the caller when the scheduler drops a cancell |
|
Concise, domain-specific, not formal or |
2026-07-05 |
| pr |
0.05 |
Modify Vulkan SDK installation command |
|
Brief update, no polite or generic AI ph |
2026-07-05 |
| pr |
0.05 |
cuda: add CC 10.0 for linux in CUDA v12 |
|
Brief, technical commit message; project |
2026-07-04 |
| pr |
0.05 |
server: report the same capabilities in /api/tags as /api/sh |
|
Domain-specific, informal, technical sum |
2026-07-02 |
| pr |
0.05 |
agent: harness core |
|
Structured but concise; no AI hallmarks |
2026-06-29 |
| pr |
0.05 |
discover: fall back to standard CUDA when the JetPack runner |
|
Technical, direct, and responds to a hum |
2026-06-28 |
| pr |
0.05 |
Add model eval metrics to /metrics |
|
Contains direct metric examples, minimal |
2025-06-22 |
| pr |
0.05 |
app/ui: allow pasting images and files from the clipboard in |
|
Uses a PR template; free-text is concise |
2026-04-13 |
| pr |
0.05 |
Include parser capabilities in model list |
|
Template used, but free-text is detailed |
2026-06-30 |
| pr |
0.05 |
docs: fix incorrect tr command in Linux uninstall instructio |
|
Technical, informal, reviewer comment; c |
2026-03-18 |
| pr |
0.05 |
fix: correctly count CRLF line endings in modelfile parser |
|
Terse, technical changelog, reviewer com |
2026-03-18 |
| pr |
0.05 |
fix(app): remove sidebar open animation on initial load |
|
Structured, but all custom and technical |
2026-02-28 |
| pr |
0.05 |
llm: warn on large CUDA host model buffers |
|
Highly technical, domain-heavy, not AI-g |
2026-07-01 |
| pr |
0.05 |
openai: respect explicit top_p=0 in legacy completions |
|
Detailed, technical, compares pointer se |
2026-07-01 |
| pr |
0.05 |
server: add cached eval metric to response |
|
Technical, concise, with changelog/issue |
2026-06-25 |
| pr |
0.05 |
cmd/interactive.go: convert "true"/"false" strings to bools |
|
Issue/fix breakdown in technical changel |
2026-06-13 |
| pr |
0.05 |
cmd: avoid terminal wrapping when stdout is redirected |
|
Domain-specific, concise explanations; n |
2026-06-29 |
| pr |
0.05 |
launch: check for min version for hermes desktop |
|
Natural developer tone, references, and |
2026-06-25 |
| pr |
0.05 |
server: inherit context_length and base name from parent mo |
|
Partial word at end suggests human editi |
2026-06-29 |
| pr |
0.03 |
mlx: bump dependency |
|
Standard terse commit message plus techn |
2026-06-27 |
| pr |
0.01 |
docs: Redesign OpenClaw integration page |
|
Template-only, no free-form content to a |
2026-06-24 |
| pr |
0.01 |
feat: add persian documentation and windows setup script |
|
Minimal info, abrupt—typical terse human |
2026-07-02 |
| pr |
0.01 |
docs: update vscode docs |
|
Brief summary, template use—no AI signal |
2026-06-25 |
| pr |
0.01 |
docs(cloud): update retirement list |
|
Template-only, no free-form content ente |
2026-07-02 |
| pr |
0.01 |
llama.cpp update |
|
Very brief, lacking AI-generic phrasing |
2026-06-29 |
| commit |
0.00 |
llama: enable FA on CUDA CC 6.x GPUs (#16994) |
|
Technical language with domain specifics |
2026-07-03 |
| commit |
0.00 |
compat: use UTF-8-safe file open (#16999) |
|
Brief, technical summary with human-like |
2026-07-02 |
| commit |
0.00 |
rocm: remove no longer supported devices (#17010) |
|
Informal, human tone and concise problem |
2026-07-02 |
| commit |
0.00 |
llama: clean up dead code from llama-server work (#17007) |
|
Terse, technical language; typical human |
2026-07-02 |
| commit |
0.00 |
agent: harness core (#16963) |
|
Minimal, standard changelog line with no |
2026-07-02 |
| commit |
0.00 |
discover: fall back to standard CUDA when the JetPack runner |
|
Extensive technical detail, domain jargo |
2026-07-02 |
| commit |
0.00 |
docs(cloud): update retirement list (#17000) |
|
Standard doc update, no free-text or AI |
2026-07-02 |
| commit |
0.00 |
fix cuda toolkit lookup and parallel (#16613) |
|
Terse commit messages with technical jar |
2026-06-30 |
| commit |
0.00 |
ci: avoid unbounded parallelism (#16966) |
|
Casual tone, domain-specific explanation |
2026-06-30 |
| commit |
0.00 |
mlx: tighten up gemma4 moe loading code (#16964) |
|
Technical content with human phrasing an |
2026-06-30 |
| commit |
0.00 |
launch: check for min version for hermes desktop (#16912) |
|
Terse commit message, technical, no AI s |
2026-06-29 |
| commit |
0.00 |
llama.cpp update (#16960) |
|
Brief update message, no AI characterist |
2026-06-29 |
| commit |
0.00 |
mlx: bump dependency (#16935) |
|
Contains tech jargon, concise and direct |
2026-06-29 |
| commit |
0.00 |
tools: ignore braces inside JSON strings when detecting tool |
|
Concise, technical language with some in |
2026-06-27 |
| commit |
0.00 |
parser/renderer: add Ornith 9B renderer/parser support (#169 |
|
Very terse, uses minimal language; human |
2026-06-26 |
| commit |
0.00 |
launch: update hermes install urls to official (#16913) |
|
Concise, technical commit, typical human |
2026-06-25 |
| commit |
0.00 |
fix capability grid dark mode style (#16907) |
|
Terse bugfix commit, minimal context, cl |
2026-06-25 |
| commit |
0.00 |
CUDA: require driver 550 or newer for v12 (#16895) |
|
Contains technical jargon, details, huma |
2026-06-25 |
| commit |
0.00 |
docs: redesign coding integration docs (#16808) |
|
Very brief, typical for docs update comm |
2026-06-25 |
| commit |
0.00 |
llama.cpp version update (#16548) |
|
Sparse, standard technical commit title, |
2026-06-24 |
| commit |
0.00 |
jetson: add CC 87 for CUDA v13 (#16628) |
|
Concise title and brief technical explan |
2026-06-24 |
| commit |
0.00 |
server: align generate with native chat templates (#16878) |
|
Technical jargon, informal phrasing, no |
2026-06-24 |
| commit |
0.00 |
docs: redesign docs landing and integrations overview (#1680 |
|
Standard docs commit message with co-aut |
2026-06-24 |
| commit |
0.00 |
llm: fix ollama ps double-counting mmap'd weights on partial |
|
Detailed technical explanation, normal i |
2026-06-24 |
| commit |
0.00 |
mlx: update and fix CUDA JIT packaging (#16871) |
|
Technical, specific, human-like explanat |
2026-06-24 |
| commit |
0.00 |
llm: use host Vulkan loader on Windows (#16869) |
|
Short technical summary, human phrasing. |
2026-06-24 |
| commit |
0.00 |
llama: default qwen2.5vl window attention metadata (#16868) |
|
Detailed domain-specific description, in |
2026-06-24 |
| commit |
0.00 |
llm: preserve generation headroom for shifted prompts (#1685 |
|
Brief message, touches on context, no AI |
2026-06-23 |
| commit |
0.00 |
docs: document max think level (#16877) |
|
Concise, uses common docs commit format, |
2026-06-23 |
| commit |
0.00 |
llm: size mmproj offload by projector memory (#16866) |
|
Technical, domain acronyms, no generic o |
2026-06-23 |
| commit |
0.00 |
llama: add sm_86 architecture to cuda_v13_windows preset (#1 |
|
Detailed technical explanation, signed b |
2026-06-23 |
| commit |
0.00 |
launch/codex: detect model drift when Codex App UI switches |
|
Concise, technical summary with context- |
2026-06-22 |
| commit |
0.00 |
mlxrunner: choose the speculative draft length to maximize t |
|
Uses technical abbreviations and concise |
2026-06-10 |
| commit |
0.00 |
mlxrunner: resolve each speculative round in one host sync |
|
Efficient phrasing and technical descrip |
2026-06-11 |
| commit |
0.00 |
mlxrunner: run one target forward per MTP decode step |
|
Terse, technical, and uses inline explan |
2026-06-10 |
| commit |
0.00 |
mlxrunner: apply in-flight drafts to proposal penalty histor |
|
Precise technical explanation, abbreviat |
2026-06-12 |
| commit |
0.00 |
mlxrunner: support draft heads that maintain draft caches |
|
Domain-specific language, clear technica |
2026-06-10 |
| commit |
0.00 |
mlxrunner: host speculative decoding in the text generation |
|
Direct technical restructuring, no boile |
2026-06-09 |
| commit |
0.00 |
mlxrunner: unify the MTP decode paths |
|
Concise explanation of code changes, ver |
2026-06-02 |
| commit |
0.00 |
discover: fix inverted iGPU/dGPU Vulkan classification on Wi |
|
Specific bug description, technical deta |
2026-06-22 |
| commit |
0.00 |
launch: add thinking capability detection to opencode (#1543 |
|
Commit message is terse and technical wi |
2026-06-18 |
| commit |
0.00 |
ci: pin darwin release xcode (#16788) |
|
Terse commit message, no AI signals. |
2026-06-17 |
| commit |
0.00 |
llama: update llama.cpp to b9672 (#16775) |
|
Brief update, domain specific, human-lik |
2026-06-17 |
| commit |
0.00 |
models: add cohere2_moe (Command A / North) to the MLX engin |
|
Includes jargon, succinct, typical human |
2026-06-17 |
| commit |
0.00 |
llm: context shift allow shiftable prompts (#16764) |
|
Terse commit title, zero AI signals. |
2026-06-16 |
| commit |
0.00 |
integration: look for the "hf" tool in integration tests (#1 |
|
Concise, technical, no AI hallmarks. |
2026-06-16 |
| commit |
0.00 |
launch: Fix launch provider drift (#16683) |
|
Brief message with domain jargon and inf |
2026-06-12 |
| commit |
0.00 |
mlxrunner: record committed MTP drafts before streaming them |
|
Technical, terse, contains domain-specif |
2026-06-01 |
| commit |
0.00 |
mlxrunner: capture prefill snapshots across the forward |
|
Contains domain-specific terms and natur |
2026-05-29 |
| commit |
0.00 |
mlxrunner: drive MTP speculation through cache snapshots |
|
Casual engineering tone, with technical |
2026-05-29 |
| commit |
0.00 |
nn/recurrent: return per-boundary states from the gated-delt |
|
Technical with specific kernel/layer det |
2026-05-29 |
| commit |
0.00 |
mlxrunner/cache: split KVCache and RotatingKVCache into thei |
|
File reorganization with clear, brief en |
2026-05-29 |
| commit |
0.00 |
mlxthread: preserve the original stack when worker work pani |
|
Precise, technical description with natu |
2026-05-27 |
| commit |
0.00 |
docs: omp (#16552) |
|
Terse, domain-specific commit; no AI sig |
2026-06-08 |
| commit |
0.00 |
docs: hermes desktop (#16549) |
|
Minimal content; no AI stylistic feature |
2026-06-08 |
| commit |
0.00 |
docs: update docs examples to use Gemma 4 instead of Gemma 3 |
|
Standard technical commit message, no AI |
2026-06-07 |
| commit |
0.00 |
Add AGENTS.md and CLAUDE.md to root repository (#16604) |
|
Concise subject, no AI signals present. |
2026-06-07 |
| commit |
0.00 |
docs: describe llama.cpp update process (#16603) |
|
Standard technical commit summary, human |
2026-06-07 |
| commit |
0.00 |
openai: align models list with tags (#16556) |
|
Concise commit message, domain shorthand |
2026-06-06 |
| commit |
0.00 |
launch: use native Windows Hermes config path (#16558) |
|
Brief technical message, human style, no |
2026-06-06 |
| commit |
0.00 |
docs: update Zod example to use native toJSONSchema (#14746) |
|
Short technical message, human, includes |
2026-06-05 |
| commit |
0.00 |
launch: oh-my-pi (#16410) |
|
Commit message is brief and typical of h |
2026-06-05 |
| commit |
0.00 |
mlx: allow the embedding layer to use the nvfp4 global scale |
|
Technical title, no signs of AI phrasing |
2026-06-05 |
| commit |
0.00 |
docs: add cloud model retirement (#16528) |
|
Minimal changelog, concise and domain-sp |
2026-06-04 |
| commit |
0.00 |
launch: hermes-desktop app (#16516) |
|
Direct and brief, uses domain language a |
2026-06-04 |
| commit |
0.00 |
llama.cpp version update (#16511) |
|
Detailed but technical, with issue refer |
2026-06-04 |
| commit |
0.00 |
integrations: hermes windows install (#16487) |
|
Very brief changelog, standard human com |
2026-06-04 |
| commit |
0.00 |
docs: add Cline CLI integration doc (#16341) |
|
Very short, docs update typical for huma |
2026-06-04 |
| commit |
0.00 |
llama-server: fix gemma4 patch wiring (#16477) |
|
Terse technical commit message with spec |
2026-06-03 |
| commit |
0.00 |
launch: migrate pi (#16213) |
|
Brief, informal, just a technical migrat |
2026-06-03 |
| commit |
0.00 |
llama-server: allow GPU offload for projectors (#16473) |
|
Contains jargon and specific details; in |
2026-06-03 |
| commit |
0.00 |
server: classify mmproj GGUFs as projector layers (#16472) |
|
Direct technical commit; no AI signals. |
2026-06-03 |
| commit |
0.00 |
launch: clean legacy codex profile before launch (#16467) |
|
Short, domain-specific; no sign of AI ge |
2026-06-03 |
| commit |
0.00 |
Kill llama-server during Windows cleanup (#16458) |
|
Clear, specific description with informa |
2026-06-03 |
| commit |
0.00 |
llama.cpp version update (#16463) |
|
Technical changelog; concise and domain- |
2026-06-03 |
| commit |
0.00 |
models: add support for gemma4-12b (#16457) |
|
Very terse and standard technical commit |
2026-06-03 |
| commit |
0.00 |
fix laguna patch build breakage (#16445) |
|
Technical fix with domain details and te |
2026-06-02 |
| commit |
0.00 |
llm: ignore llama-server SSE ping comments (#16443) |
|
Contains domain jargon, informal explana |
2026-06-02 |
| commit |
0.00 |
docs: configure hermes desktop app (#16440) |
|
Minimal template-driven docs commit, no |
2026-06-02 |
| commit |
0.00 |
llama: add laguna (poolside) arch via a llama.cpp patch unde |
|
Contains technical specifics and human c |
2026-06-02 |
| commit |
0.00 |
launch: isolate Codex launch configuration (#16437) |
|
Short, domain-specific commit message; n |
2026-06-02 |
| commit |
0.00 |
llama.cpp version update (#16426) |
|
Generic update message, but no AI marker |
2026-06-02 |
| commit |
0.00 |
More harden app markdown URL handling (#16436) |
|
Domain phrase, human-typical wording for |
2026-06-02 |
| commit |
0.00 |
llm: detect llama-server load stalls from output (#16427) |
|
Complex technical description with jargo |
2026-06-02 |
| commit |
0.00 |
discover: allow Radeon 8060S iGPU by default (#16429) |
|
Detailed technical explanation, issue re |
2026-06-02 |
| commit |
0.00 |
Harden app markdown URL handling (#16380) |
|
Human-style commit with domain-specific |
2026-06-02 |
| commit |
0.00 |
llm: include cached prompt tokens in llama-server counts (#1 |
|
Contains technical jargon and concise ex |
2026-06-02 |
| commit |
0.00 |
launch: fix opencode local model limits (#16425) |
|
Technical language, no signs of AI-style |
2026-06-02 |
| commit |
0.00 |
cmd/launch: add Qwen code integration (#15900) |
|
Minimal message, no AI indicators. |
2026-06-01 |
| pr |
0.00 |
ggml: add loongarch quantization source |
|
Brief, technical changes, clear test ste |
2026-05-15 |
| pr |
0.00 |
docs: document Linux context length service config |
|
Template structure, short edits, terse a |
2026-05-15 |
| pr |
0.00 |
docs: fix GPU docs FAQ link |
|
Very brief, domain appropriate, no AI si |
2026-05-15 |
| pr |
0.00 |
test: fix data race in unit tests |
|
Short, to-the-point commit message; uses |
2025-11-18 |
| pr |
0.00 |
anthropic: include empty input for parameterless tool calls |
|
Uses detailed technical references and n |
2026-07-03 |
| pr |
0.00 |
ci: wire up MLX unit tests for PR runs |
|
Informal style, context notes, no AI hal |
2026-07-03 |
| pr |
0.00 |
test: revamp integration test entrypoints |
|
Technical, direct editorial phrasing and |
2026-06-06 |
| pr |
0.00 |
parsers: track nesting depth when closing qwen3-coder tool c |
|
Specific code logic and concise technica |
2026-07-03 |
| pr |
0.00 |
readme: add OpenAgentic to Frameworks & Agents |
|
Detailed with project-specific details, |
2026-06-25 |
| pr |
0.00 |
create: drop Qwen MTP tensors during MLX import |
|
Uses technical jargon and includes a cop |
2026-07-03 |
| pr |
0.00 |
Add agent TUI |
|
Structured technical changelog, concise |
2026-07-03 |
| pr |
0.00 |
win: support CUDA on Windows ARM64 |
|
Terse, one-line draft note; tone is huma |
2026-06-26 |
| pr |
0.00 |
rocm: remove no longer supported devices |
|
Technical explanation, includes minor ty |
2026-07-02 |
| pr |
0.00 |
server: reduce automatic context for large model fits |
|
Detailed, domain-specific, human technic |
2026-07-01 |
| pr |
0.00 |
server: preserve parser tokens for raw generate |
|
Uses context, domain-specific language, |
2026-06-30 |
| pr |
0.00 |
server: fix unified Gemma4 vision support |
|
Technical, domain-rich explanations, cle |
2026-06-23 |
| pr |
0.00 |
llama: clean up dead code from llama-server work |
|
Brief technical summary, no AI-style lan |
2026-07-02 |
| pr |
0.00 |
mlx: follow up for gemma4 |
|
Terse reply; standard human follow-up ph |
2026-04-13 |
| pr |
0.00 |
model/parsers: honor think value when initializing the qwen3 |
|
Detailed, domain-specific explanation wi |
2026-06-29 |
| pr |
0.00 |
sched: align placement reserve with llama fit target |
|
Highly technical, uses domain-specific t |
2026-06-24 |
| pr |
0.00 |
ci: avoid unbounded parallelism |
|
Technical context, timeline and numbers, |
2026-06-30 |
| pr |
0.00 |
server: remove OLLAMA_EXPERIMENT=client2 |
|
Short rationale, direct and domain-speci |
2026-06-29 |
| pr |
0.00 |
mlx: tighten up gemma4 moe loading code |
|
Terse, technical explanation with no AI |
2026-06-30 |
| pr |
0.00 |
app/tools: keep web search snippets from splitting UTF-8 cha |
|
Contains inline code and informal, techn |
2026-06-30 |
| pr |
0.00 |
launch: add Darwin MLX recommendations |
|
Uses domain-specific terms, bullet forma |
2026-06-17 |
| pr |
0.00 |
docs: add community integration to README |
|
Informal tone, natural language, inconsi |
2026-03-24 |
| pr |
0.00 |
server: truncate embedding inputs to the physical batch size |
|
Detailed bug explanation, with domain-sp |
2026-06-12 |
| pr |
0.00 |
llm: skip quantized kv cache for models that don't support i |
|
Bug explanation with brief references an |
2026-06-29 |
| pr |
0.00 |
docs: update hermes docs |
|
Generic template header, no free-text co |
2026-06-25 |
| pr |
0.00 |
app/ui: render chat messages with automatic text direction |
|
Clear, specific explanation and technica |
2026-06-29 |
| pr |
0.00 |
fix: resolve Gemma4 QAT vision hang and Jetson projector off |
|
Dense with domain-specific jargon and in |
2026-06-29 |
| pr |
0.00 |
discover: recognize the SYCL ggml backend |
|
Uses technical details, personal testing |
2026-06-27 |
| pr |
0.00 |
llm: disable llama fit for Gemma4 on limited VRAM |
|
Terse, technical bullet list and command |
2026-06-20 |
| pr |
0.00 |
llm: size mmproj offload by projector |
|
Highly technical, references specific fa |
2026-06-15 |
