{"id":46535341,"url":"https://github.com/johnmai-dev/ane-lm","last_synced_at":"2026-03-10T02:01:02.660Z","repository":{"id":341896216,"uuid":"1171731709","full_name":"johnmai-dev/ANE-LM","owner":"johnmai-dev","description":"LLM inference on Apple Neural Engine (ANE) ","archived":false,"fork":false,"pushed_at":"2026-03-04T11:31:14.000Z","size":163,"stargazers_count":115,"open_issues_count":3,"forks_count":8,"subscribers_count":4,"default_branch":"main","last_synced_at":"2026-03-09T03:57:37.890Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":"","language":"C++","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/johnmai-dev.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2026-03-03T14:50:26.000Z","updated_at":"2026-03-08T07:34:34.000Z","dependencies_parsed_at":"2026-03-08T00:00:46.647Z","dependency_job_id":null,"html_url":"https://github.com/johnmai-dev/ANE-LM","commit_stats":null,"previous_names":["johnmai-dev/ane-lm"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/johnmai-dev/ANE-LM","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/johnmai-dev%2FANE-LM","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/johnmai-dev%2FANE-LM/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/johnmai-dev%2FANE-LM/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/johnmai-dev%2FANE-LM/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/johnmai-dev","download_url":"https://codeload.github.com/johnmai-dev/ANE-LM/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/johnmai-dev%2FANE-LM/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":30322634,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-03-10T01:36:58.598Z","status":"online","status_checked_at":"2026-03-10T02:00:06.579Z","response_time":106,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":[],"created_at":"2026-03-06T23:33:05.821Z","updated_at":"2026-03-10T02:01:02.600Z","avatar_url":"https://github.com/johnmai-dev.png","language":"C++","funding_links":[],"categories":[],"sub_categories":[],"readme":"# ANE-LM\n\nLLM inference on Apple Neural Engine (ANE) using private `AppleNeuralEngine.framework` APIs. \n## Supported Models\n\n- Qwen3 (dense)\n- Qwen3.5 (dense, text-only)\n\n## Build\n\n```bash\ncmake -B build -DCMAKE_BUILD_TYPE=Release\ncmake --build build\n```\n\n## Usage\n\n![image](assets/image.png)\n\nDownload a supported model (e.g. `Qwen3-0.6B` or `Qwen3.5-0.8B` in safetensors format), then:\n\n```bash\n# Single-shot generation\n./build/ane-lm generate --model /path/to/Qwen3.5-0.8B --prompt \"Hello\"\n\n# Interactive chat\n./build/ane-lm chat --model /path/to/Qwen3.5-0.8B\n\n# Pre-convert weights (BF16 -\u003e FP16, speeds up subsequent loads)\n./build/ane-lm convert --model /path/to/Qwen3.5-0.8B\n```\n\n### Options\n\n```\n--model \u003cpath\u003e       Path to model directory (required)\n--prompt \u003ctext\u003e      Input prompt (generate mode, default: \"Hello\")\n--max-tokens N       Max tokens to generate (default: unlimited)\n--temp T             Temperature (default: 0.6)\n--repeat-penalty P   Repetition penalty (default: 1.2, 1.0=off)\n--enable-thinking    Enable thinking/reasoning mode\n--no-ane-cache       Disable persistent ANE compile cache\n-v, --verbose        Show detailed initialization info\n```\n\n## Requirements\n\n- macOS 13.0+\n- Apple Silicon (M1/M2/M3/M4/M5)\n\n## Acknowledgments\n\n- [maderix/ANE](https://github.com/maderix/ANE) - Training neural networks on Apple Neural Engine via reverse-engineered private APIs\n- [llama.cpp](https://github.com/ggml-org/llama.cpp) - LLM inference in C/C++\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fjohnmai-dev%2Fane-lm","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fjohnmai-dev%2Fane-lm","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fjohnmai-dev%2Fane-lm/lists"}