{"slug":"chubbyguan-podcast-transcribe-04fbcc","title":"\u64ad\u5ba2\u4e0b\u8f7d\u4e0e\u8f6c\u5f55\u5de5\u4f5c\u6d41","description":"\u5c06\u5c0f\u5b87\u5b99\u3001\u559c\u9a6c\u62c9\u96c5\u3001\u76f4\u94fe\u97f3\u9891\u6216\u672c\u5730\u97f3\u9891\u4e0b\u8f7d\u5e76\u8f6c\u5f55\u4e3a\u5e26 frontmatter \u7684 Markdown\uff0c\u652f\u6301 RSS \u6279\u91cf\u3001\u672c\u5730 SenseVoice \u4e0e\u53ef\u9009\u4e91\u7aef DashScope/Groq \u540e\u7aef\u3001\u957f\u97f3\u9891\u5206\u7247\u7eed\u4f20\u548c\u4e91\u7aef\u4efb\u52a1\u72b6\u6001\u6062\u590d\u3002\u6587\u6863\u5bf9\u9690\u79c1\u3001\u8ba1\u8d39\u4e0e\u8bc6\u522b\u8bef\u5dee\u6709\u660e\u786e\u8fb9\u754c\u8bf4\u660e\u3002","category":"\u9009\u9898\u4e0e\u7814\u7a76","tags":["chubbyguan","\u9009\u9898\u4e0e\u7814\u7a76"],"author":"chubbyguan","source_url":"https://github.com/chubbyguan/chubbyskills/blob/d066d7694d23045a39f8c3efc1b51ebb9ca7f867/podcast-transcribe/SKILL.md","license":"MIT","status":"published","verification":"source_reviewed","verification_notes":"\u5df2\u6838\u5bf9\u539f\u59cb\u4ed3\u5e93\u3001\u56fa\u5b9a\u63d0\u4ea4\u3001\u8bb8\u53ef\u3001\u6280\u80fd\u6b63\u6587\u4e0e\u5305\u7ed3\u6784\uff0c\u5e76\u8865\u5145\u4e2d\u6587\u6267\u884c\u9a8c\u6536\u8bf4\u660e\uff1b\u672a\u8fd0\u884c\u5305\u5185\u811a\u672c\u3002","revision":"8eeaa1089aea428b8aa0854c6172dcde","created_at":"2026-11-02T14:47:08+00:00","updated_at":"2026-11-02T14:47:08+00:00","source_repo":"chubbyguan/chubbyskills","source_path":"podcast-transcribe/SKILL.md","source_commit":"d066d7694d23045a39f8c3efc1b51ebb9ca7f867","source_updated_at":"2026-11-02T14:46:04+00:00","stars":0,"proxy_score":null,"score":null,"grade":null,"dimensions":{},"evaluation":{},"installs":0,"platforms":[],"formats":[],"category_slug":"research","category_name":"\u9009\u9898\u4e0e\u7814\u7a76","quality":{"review_type":"text_and_package_review","readiness":"requires_setup","reviewed_at":"2026-11-02T14:46:04+00:00","review_summary":"\u5de5\u4f5c\u6d41\u5b8c\u6574\u3001\u540e\u7aef\u9009\u62e9\u4e0e\u9690\u79c1\u8ba1\u8d39\u8fb9\u754c\u8bf4\u660e\u6e05\u695a\uff0c\u4f46\u4f9d\u8d56\u5916\u90e8\u5171\u7528\u6a21\u5757\u4e0e\u6a21\u578b\u4e0b\u8f7d\uff0c\u9700\u914d\u7f6e\u540e\u4f7f\u7528\uff1b\u811a\u672c\u672a\u8fd0\u884c\u9a8c\u8bc1\u3002","cost_notes":"\u4e91\u7aef\u540e\u7aef\uff08DashScope\u3001Groq\uff09\u6309\u670d\u52a1\u5546\u8ba1\u8d39\uff0cGroq \u514d\u8d39\u5c42\u6709\u5355\u6587\u4ef6 25MB \u4e0a\u9650\uff1b\u5177\u4f53\u4ef7\u683c\u4e0e\u989d\u5ea6\u4ee5\u670d\u52a1\u5546\u516c\u793a\u4e3a\u51c6\uff0c\u9700\u7528\u6237\u81ea\u884c\u786e\u8ba4\u3002\u672c\u5730\u540e\u7aef\u65e0 API \u8d39\u7528\u4f46\u6d88\u8017\u672c\u673a\u7b97\u529b\u3002","source_sha256":"f4f037ca5a66febd4f9ef4bf97d01eb2d968b6375aafecd7880f5c3addfa3dcd","prerequisites":["Python 3.11 \u6216\u66f4\u65b0\u7248\u672c\u53ca requirements.txt \u4e2d\u7684\u4f9d\u8d56","ffmpeg\uff08macOS \u7528 brew\uff0cUbuntu \u7528 apt \u5b89\u88c5\uff09","\u672c\u5730\u540e\u7aef\u9700\u4e0b\u8f7d SenseVoice-Small \u6a21\u578b\uff1b\u53ef\u9009 qwen3-asr-0.6b \u9700\u53e6\u88c5 qwen-asr\u3001transformers\u3001torch","\u4e91\u7aef\u540e\u7aef\u9700 DASHSCOPE_API_KEY \u6216 GROQ_API_KEY\uff0c\u901a\u8fc7\u5b89\u5168\u73af\u5883\u914d\u7f6e","\u7f51\u7edc\u8bbf\u95ee\u97f3\u9891\u6765\u6e90\u4e0e\uff08\u5982\u4f7f\u7528\uff09\u4e91\u7aef\u8f6c\u5f55\u670d\u52a1"],"checks":["\u751f\u6210\u7684 Markdown \u542b frontmatter\uff0c\u6765\u6e90\u94fe\u63a5\u4e0e\u6240\u7528\u540e\u7aef\u6807\u8bb0\u6b63\u786e","\u65e0\u5206\u6bb5\u4fe1\u606f\u65f6\u6587\u7a3f\u4e0d\u542b\u7f16\u9020\u7684\u65f6\u95f4\u6233","Groq \u957f\u97f3\u9891\u5206\u7247\u540e\uff0c\u5404\u6bb5\u65f6\u95f4\u6233\u6309\u987a\u5e8f\u7d2f\u52a0\u504f\u79fb\uff0c\u65e0\u91cd\u590d\u6216\u8df3\u6bb5","\u4e2d\u65ad\u540e\u91cd\u8dd1\u540c\u4e00\u547d\u4ee4\u80fd\u590d\u7528\u72b6\u6001\u76ee\u5f55\uff0c\u4e0d\u91cd\u590d\u63d0\u4ea4\u5df2\u5b8c\u6210\u5206\u7247","\u81ea\u52a8\u4e0b\u8f7d\u4ec5\u63a5\u53d7\u516c\u7f51 HTTP(S) \u76f4\u8fde\uff0c\u62d2\u7edd\u672c\u5730\u6216\u79c1\u7f51\u5730\u5740"],"limitations":["\u672c\u5730 SenseVoice-Small \u53ea\u8f93\u51fa\u7eaf\u6587\u672c\u5168\u6587\uff0c\u4e0d\u751f\u6210\u9010\u6bb5\u65f6\u95f4\u6233","\u8f6c\u5f55\u53ef\u80fd\u5b58\u5728\u4e13\u6709\u540d\u8bcd\u3001\u6570\u5b57\u6216\u65ad\u53e5\u9519\u8bef\uff0c\u5f15\u7528\u524d\u9700\u6838\u5bf9\u539f\u59cb\u97f3\u9891","\u4e0d\u63d0\u4f9b\u901a\u7528\u8bf4\u8bdd\u4eba\u5206\u79bb\u4fdd\u8bc1","\u4e91\u7aef\u771f\u5b9e\u53ef\u7528\u6027\u3001\u8d26\u53f7\u6743\u9650\u3001\u97f3\u9891\u517c\u5bb9\u6027\u4e0e\u8d39\u7528\u5c1a\u672a\u72ec\u7acb\u9a8c\u6536\uff1b\u672c\u6b21\u4ec5\u505a\u6587\u672c\u4e0e\u5305\u7ed3\u6784\u5ba1\u9605\uff0c\u672a\u8fd0\u884c\u811a\u672c"],"issues":["SKILL.md \u63d0\u5230\u5171\u7528 chubby_common/funasr.py \u5c01\u88c5\uff0c\u8be5\u6587\u4ef6\u4e0d\u5728\u672c\u5305\u6587\u4ef6\u6e05\u5355\u5185\uff0c\u672c\u5730\u540e\u7aef\u80fd\u5426\u76f4\u63a5\u8fd0\u884c\u5f85\u6838\u5b9e","\u6587\u6863\u4e2d\u7684\u901f\u5ea6\u5bf9\u6bd4\u6570\u636e\uff08RTF 0.11 \u4e0e 0.80\uff09\u6765\u81ea\u4f5c\u8005\u81ea\u8ff0\u5b9e\u6d4b\uff0c\u672c\u5ba1\u9605\u672a\u590d\u73b0\uff0c\u7528\u6237\u4e0d\u5e94\u636e\u6b64\u4f5c\u6027\u80fd\u627f\u8bfa","\u9700\u8981\u8df3\u8f6c\u6216\u4ee3\u7406\u7684\u97f3\u9891\u6765\u6e90\u4e0d\u88ab\u81ea\u52a8\u4e0b\u8f7d\u652f\u6301\uff0c\u9700\u7528\u6237\u5148\u81ea\u884c\u4e0b\u8f7d"],"missing_files":[]},"quality_url":"https://shiliuai.cn/api/v2/skills/chubbyguan-podcast-transcribe-04fbcc/files/SHILIU-QUALITY.md","usage_guide":{"task":"\u628a\u5355\u96c6\u64ad\u5ba2\u94fe\u63a5\u3001RSS \u8ba2\u9605\u6216\u672c\u5730\u97f3\u9891\u6587\u4ef6\u8f6c\u5f55\u4e3a Markdown \u6587\u7a3f","inputs":"\u64ad\u5ba2\u5355\u96c6\u9875\u9762\u94fe\u63a5\u3001RSS \u5730\u5740\u6216\u672c\u5730\u97f3\u9891\u6587\u4ef6\u8def\u5f84\uff1bPython 3.11+ \u4e0e requirements.txt \u4f9d\u8d56\uff1bffmpeg\uff08\u5bb9\u5668\u8f6c\u6362\u4e0e\u5206\u7247\u9700\u8981\uff09\uff1b\u672c\u5730\u540e\u7aef\u9996\u6b21\u8fd0\u884c\u9700\u4e0b\u8f7d SenseVoice \u6a21\u578b\uff1b\u4f7f\u7528\u4e91\u7aef\u540e\u7aef\u65f6\u9700 DASHSCOPE_API_KEY \u6216 GROQ_API_KEY\uff0c\u901a\u8fc7\u73af\u5883\u53d8\u91cf\u914d\u7f6e","outputs":"\u5e26 frontmatter\u3001\u6765\u6e90\u94fe\u63a5\u548c\u8f6c\u5f55\u540e\u7aef\u6807\u8bb0\u7684 Markdown \u6587\u7a3f\uff1b\u540e\u7aef\u8fd4\u56de\u5206\u6bb5\u65f6\u4fdd\u7559\u65f6\u95f4\u6233\uff0cGroq \u5206\u7247\u65f6\u9644\u5206\u6bb5\u65f6\u95f4\u6233\u9644\u5f55\uff1b\u6807\u51c6\u8f93\u51fa\u6700\u540e\u4e00\u884c\u4e3a\u751f\u6210\u6587\u4ef6\u8def\u5f84","example_input":"\u8bf7\u7528 podcast-transcribe \u8f6c\u5f55\u8fd9\u4e2a\u672c\u5730\u97f3\u9891 /data/episodes/ep01.mp3\uff0c\u4f7f\u7528\u672c\u5730 SenseVoice \u540e\u7aef\uff0c\u8f93\u51fa\u5230 ./output\u3002","example_output":"\u793a\u610f\uff1a./output/ep01.md\uff0c\u542b frontmatter\uff08\u6807\u9898\u3001\u6765\u6e90\u3001\u540e\u7aef\u6807\u8bb0\uff09\u4e0e\u8f6c\u5f55\u6b63\u6587\uff1b\u82e5\u540e\u7aef\u65e0\u5206\u6bb5\u4fe1\u606f\u5219\u53ea\u6709\u5168\u6587\uff0c\u4e0d\u751f\u6210\u65f6\u95f4\u8f74\u3002\u6b64\u5904\u4e0d\u5c55\u793a\u771f\u5b9e\u8f6c\u5f55\u5185\u5bb9\u3002"},"read_url":"https://shiliuai.cn/api/v2/skills/chubbyguan-podcast-transcribe-04fbcc","skill_url":"https://shiliuai.cn/api/v2/skills/chubbyguan-podcast-transcribe-04fbcc/files/SKILL.md","download_url":"https://shiliuai.cn/api/v2/skills/chubbyguan-podcast-transcribe-04fbcc/download","install_url":"https://shiliuai.cn/api/v2/skills/chubbyguan-podcast-transcribe-04fbcc/install","files":[{"path":"LICENSE","bytes":1063,"url":"https://shiliuai.cn/api/v2/skills/chubbyguan-podcast-transcribe-04fbcc/files/LICENSE"},{"path":"SHILIU-QUALITY.md","bytes":2286,"url":"https://shiliuai.cn/api/v2/skills/chubbyguan-podcast-transcribe-04fbcc/files/SHILIU-QUALITY.md"},{"path":"SKILL.md","bytes":7486,"url":"https://shiliuai.cn/api/v2/skills/chubbyguan-podcast-transcribe-04fbcc/files/SKILL.md"},{"path":"requirements.txt","bytes":64,"url":"https://shiliuai.cn/api/v2/skills/chubbyguan-podcast-transcribe-04fbcc/files/requirements.txt"},{"path":"scripts/batch_transcribe.py","bytes":9488,"url":"https://shiliuai.cn/api/v2/skills/chubbyguan-podcast-transcribe-04fbcc/files/scripts/batch_transcribe.py"},{"path":"scripts/cloud_transcribe.py","bytes":27598,"url":"https://shiliuai.cn/api/v2/skills/chubbyguan-podcast-transcribe-04fbcc/files/scripts/cloud_transcribe.py"},{"path":"scripts/local_qwen_asr.py","bytes":1847,"url":"https://shiliuai.cn/api/v2/skills/chubbyguan-podcast-transcribe-04fbcc/files/scripts/local_qwen_asr.py"},{"path":"scripts/provider_config.py","bytes":4391,"url":"https://shiliuai.cn/api/v2/skills/chubbyguan-podcast-transcribe-04fbcc/files/scripts/provider_config.py"},{"path":"scripts/safe_download.py","bytes":7122,"url":"https://shiliuai.cn/api/v2/skills/chubbyguan-podcast-transcribe-04fbcc/files/scripts/safe_download.py"},{"path":"scripts/transcribe.py","bytes":11592,"url":"https://shiliuai.cn/api/v2/skills/chubbyguan-podcast-transcribe-04fbcc/files/scripts/transcribe.py"},{"path":"shiliu-quality.json","bytes":2173,"url":"https://shiliuai.cn/api/v2/skills/chubbyguan-podcast-transcribe-04fbcc/files/shiliu-quality.json"},{"path":"shiliu-source.json","bytes":271,"url":"https://shiliuai.cn/api/v2/skills/chubbyguan-podcast-transcribe-04fbcc/files/shiliu-source.json"},{"path":"usage.json","bytes":981,"url":"https://shiliuai.cn/api/v2/skills/chubbyguan-podcast-transcribe-04fbcc/files/usage.json"}],"content":"---\nname: podcast-transcribe\ndescription: >\n  \u64ad\u5ba2/\u5c0f\u5b87\u5b99 \u2192 \u4e0b\u8f7d \u2192 \u8f6c\u5f55 \u2192 \u5b58\u4e3a Markdown \u7684\u5b8c\u6574\u5de5\u4f5c\u6d41\u3002\n  \u652f\u6301 RSS \u6279\u91cf\u4e0b\u8f7d\u3001\u5355\u96c6\u94fe\u63a5\u8f6c\u5f55\u3002\nmetadata:\n  category: \"media\"\n  triggers: \"[\\\"\u7528\u6237\u53d1\u9001\u5c0f\u5b87\u5b99/\u64ad\u5ba2\u94fe\u63a5\\\", \\\"\u5e2e\u6211\u8f6c\u5f55\u8fd9\u4e2a\u64ad\u5ba2\\\", \\\"\u4e0b\u8f7d\u64ad\u5ba2\\\", \\\"\u6279\u91cf\u8f6c\u5f55\u64ad\u5ba2\\\"]\"\n  version: \"1.1.0\"\n  tags: \"[\\\"media\\\", \\\"audio\\\", \\\"podcast\\\", \\\"transcription\\\", \\\"xiaoyuzhou\\\"]\"\n---\n\n# \u64ad\u5ba2\u8f6c\u5f55 Skill\n\n\u5c06\u64ad\u5ba2\u97f3\u9891\u4e0b\u8f7d\u5e76\u8f6c\u5f55\u4e3a Markdown\u3002\u652f\u6301\u5c0f\u5b87\u5b99\u3001\u559c\u9a6c\u62c9\u96c5\u3001\u76f4\u63a5\u97f3\u9891\u5730\u5740\u3001\u672c\u5730\u97f3\u9891\u548c RSS \u6279\u91cf\u6d41\u7a0b\u3002\n\n\u9ed8\u8ba4\u5728\u672c\u5730\u4f7f\u7528 SenseVoice-Small\uff08\u4e0e\u89c6\u9891\u7c7b\u6280\u80fd\u5171\u7528\u7684 `chubby_common/funasr.py` \u5c01\u88c5\uff09\u3002\u53ef\u9009\u4e91\u7aef\u540e\u7aef\u4e3a\u963f\u91cc\u4e91\u767e\u70bc DashScope \u7684 `qwen3-asr-flash` \u548c Groq \u7684 `whisper-large-v3-turbo`\uff0c\u7528\u6237\u660e\u786e\u9009\u62e9\u540e\u624d\u542f\u7528\uff1b\u97f3\u9891\u4f1a\u53d1\u9001\u81f3\u4e91\u7aef\u5e76\u53ef\u80fd\u8ba1\u8d39\u3002\n\n## \u73af\u5883\u8981\u6c42\n\n\u4ee5\u4e0b\u547d\u4ee4\u5728\u672c skill \u76ee\u5f55\u8fd0\u884c\uff0c\u5efa\u8bae\u4f7f\u7528 Python 3.11 \u6216\u66f4\u65b0\u7248\u672c\uff1a\n\n```bash\npython3 -m venv .venv\nsource .venv/bin/activate\npython3 -m pip install -r requirements.txt\n# macOS: brew install ffmpeg\n# Ubuntu: sudo apt install ffmpeg\n```\n\n\u672c\u5730\u6a21\u578b\u9996\u6b21\u4f7f\u7528\u65f6\u9700\u8981\u4e0b\u8f7d\u3002\u4ec5\u4f7f\u7528\u4e91\u7aef\u540e\u7aef\u4e0d\u9700\u8981 funasr \u672c\u5730\u4f9d\u8d56\uff1b\u4e0d\u652f\u6301\u7684\u97f3\u9891\u5bb9\u5668\u8f6c\u6362\u4ecd\u53ef\u80fd\u9700\u8981 `ffmpeg`\u3002\u4e91\u7aef\u5bc6\u94a5\u901a\u8fc7\u5b89\u5168\u73af\u5883\u914d\u7f6e\uff0c\u4e0d\u80fd\u5199\u5165\u547d\u4ee4\u53c2\u6570\u3001\u8f6c\u5f55\u7a3f\u6216\u7248\u672c\u5e93\u3002\n\n## \u5355\u96c6\u4e0e\u6279\u91cf\n\n```bash\npython3 scripts/transcribe.py \"https://www.xiaoyuzhoufm.com/episode/xxxxx\" ./output \\\n  --provider local\npython3 scripts/transcribe.py \"/\u4f60\u7684\u97f3\u9891\u76ee\u5f55/episode.mp3\" ./output \\\n  --provider local --language zh\npython3 scripts/batch_transcribe.py --rss-url \"\u66ff\u6362\u4e3a\u5b9e\u9645 RSS \u5730\u5740\" \\\n  --output ./output --count 10 --provider local\n```\n\n\u5c06\u793a\u4f8b\u5730\u5740\u66ff\u6362\u4e3a\u5b9e\u9645\u6765\u6e90\u3002\u5355\u96c6\u9875\u9762\u4f1a\u5c1d\u8bd5\u63d0\u53d6\u97f3\u9891\u94fe\u63a5\uff0c\u63d0\u53d6\u5931\u8d25\u65f6\u6539\u7528\u76f4\u63a5\u97f3\u9891\u5730\u5740\u6216\u672c\u5730\u6587\u4ef6\u3002\u6807\u51c6\u8f93\u51fa\u6700\u540e\u4e00\u884c\u662f\u6210\u529f\u751f\u6210\u7684 Markdown \u8def\u5f84\u3002\n\n\u81ea\u52a8\u4e0b\u8f7d\u4ec5\u63a5\u53d7\u516c\u7f51 HTTP(S) \u76f4\u8fde\u5730\u5740\uff0c\u7981\u7528\u4ee3\u7406\u548c\u91cd\u5b9a\u5411\uff0c\u62d2\u7edd\u672c\u5730/\u79c1\u7f51\u5730\u5740\u3002\u9700\u8981\u8df3\u8f6c\u6216\u4ee3\u7406\u7684\u6765\u6e90\uff0c\u8bf7\u5148\u81ea\u884c\u4e0b\u8f7d\u97f3\u9891\uff0c\u518d\u4f20\u5165\u672c\u5730\u6587\u4ef6\u8def\u5f84\u3002\n\n## \u672c\u5730\u6a21\u578b\u9009\u62e9\uff1aSenseVoice-Small vs Qwen3-ASR-0.6B\n\n`--provider local` \u652f\u6301\u4e24\u4e2a\u672c\u5730\u6a21\u578b\uff1a`SenseVoiceSmall`\uff08\u9ed8\u8ba4\uff09\u548c `qwen3-asr-0.6b`\uff1a\n\n```bash\npython3 scripts/transcribe.py \"/\u4f60\u7684\u97f3\u9891\u76ee\u5f55/episode.mp3\" ./output --provider local\npython3 scripts/transcribe.py \"/\u4f60\u7684\u97f3\u9891\u76ee\u5f55/episode.mp3\" ./output --provider local --model qwen3-asr-0.6b\n```\n\n`qwen3-asr-0.6b` \u662f\u53ef\u9009\u91cd\u4f9d\u8d56\uff0c\u4e0d\u5728\u9ed8\u8ba4\u5b89\u88c5\u5185\uff0c\u9700\u81ea\u884c `pip install qwen-asr transformers torch`\u3002\n\n\u4ee5\u4e0b\u5bf9\u6bd4\u6570\u636e\u6765\u81ea 2026-10 \u5728 MacBook Pro\uff08Apple M3 Pro\uff0c\u7eaf CPU\uff09\u4e0a\u5bf9 5 \u5206\u949f\u4e2d\u6587\u64ad\u5ba2\u7684\u5b9e\u6d4b\uff1a\n\n| | SenseVoice-Small\uff08\u9ed8\u8ba4\uff09 | Qwen3-ASR-0.6B\uff08\u53ef\u9009\uff09 |\n|---|---|---|\n| \u901f\u5ea6 | RTF 0.11\uff085 \u5206\u949f\u97f3\u9891\u7ea6 33 \u79d2\uff09 | RTF 0.80\uff08\u7ea6 4 \u5206\u949f\uff0c\u6162\u7ea6 7 \u500d\uff09 |\n| \u4e13\u6709\u540d\u8bcd | \u4e00\u822c\uff08\"\u5ca9\u8336\"\u8bef\u4f5c\"\u76d0\u8336\"\uff0c\u4eba\u540d\u524d\u540e\u4e0d\u4e00\u81f4\uff09 | \u66f4\u7a33\uff08\"\u5ca9\u8336\"\u3001\u4eba\u540d\u8bc6\u522b\u4e00\u81f4\uff09 |\n| \u4e3b\u8981\u98ce\u9669 | \u8f93\u51fa\u6df7\u5165\u60c5\u611f\u6807\u7b7e\uff0c\u6b63\u5f0f\u6587\u7a3f\u9700\u6e05\u6d17 | \u6709\u5e7b\u89c9\u5f0f\u6539\u5199\u98ce\u9669\uff08\"\u9ec4\u91d1\u52a0\u5de5\u5382\"\u2192\"\u7687\u5e1d\u5bb6\u65cf\"\uff09\uff1b\u65e0 ITN\uff0c\u6570\u5b57\u8f93\u51fa\u4e3a\u5168\u6587\u5b57 |\n| \u957f\u97f3\u9891 | VAD \u81ea\u52a8\u5206\u6bb5\uff0c\u7a33\u5b9a | \u6574\u6bb5\u8fdb\u6a21\u578b\uff1b\u672c\u540e\u7aef\u5df2\u628a `max_new_tokens` \u8c03\u5230 4096 \u907f\u514d\u622a\u65ad |\n| \u9002\u7528\u573a\u666f | CPU \u9ed8\u8ba4\u9009\u62e9\uff0c\u957f\u64ad\u5ba2\u53cb\u597d | GPU \u673a\u5668\uff0c\u6216\u5bf9\u4e13\u540d/\u4eba\u540d\u51c6\u786e\u6027\u654f\u611f\u7684\u5185\u5bb9 |\n\n\u4e24\u8005\u8d28\u91cf\u4e92\u6709\u80dc\u8d1f\u3001\u6ca1\u6709\u4ee3\u5dee\u3002CPU \u573a\u666f\u8bf7\u4fdd\u6301\u9ed8\u8ba4 SenseVoice-Small\uff1b\u6709 GPU \u6216\u4e13\u540d\u654f\u611f\u65f6\u518d\u9009 Qwen3-ASR-0.6B\u3002\n\n## \u53ef\u9009\u4e91\u7aef\u8f6c\u5f55\n\n| \u540e\u7aef | \u51ed\u636e\u73af\u5883\u53d8\u91cf | \u9ed8\u8ba4\u6a21\u578b |\n|---|---|---|\n| `local` | \u65e0 | `SenseVoiceSmall`\uff08\u4ec5\u7528\u4e8e\u5143\u6570\u636e\u8bb0\u5f55\uff09 |\n| `dashscope` | `DASHSCOPE_API_KEY` | `qwen3-asr-flash` |\n| `groq` | `GROQ_API_KEY` | `whisper-large-v3-turbo` |\n\n\u914d\u7f6e `DASHSCOPE_API_KEY` \u6216 `GROQ_API_KEY` \u540e\uff1a\n\n```bash\npython3 scripts/transcribe.py \"/\u4f60\u7684\u97f3\u9891\u76ee\u5f55/episode.mp3\" ./output \\\n  --provider dashscope --cloud-timeout 1800 \\\n  --state-dir \"$HOME/.local/state/chubbyskills/podcast\"\npython3 scripts/transcribe.py \"/\u4f60\u7684\u97f3\u9891\u76ee\u5f55/episode.mp3\" ./output \\\n  --provider groq --cloud-timeout 1800 \\\n  --state-dir \"$HOME/.local/state/chubbyskills/podcast\"\n```\n\n`--provider` \u4f18\u5148\u4e8e `PODCAST_TRANSCRIBE_PROVIDER`\uff0c\u90fd\u672a\u6307\u5b9a\u65f6\u4f7f\u7528 `local`\u3002\u6279\u91cf\u5165\u53e3\u652f\u6301\u540c\u6837\u7684 provider\u3001\u6a21\u578b\u3001\u8bed\u8a00\u3001\u7b49\u5f85\u548c\u72b6\u6001\u76ee\u5f55\u53c2\u6570\uff0c\u5e76\u628a\u6700\u7ec8\u9009\u9879\u663e\u5f0f\u4f20\u7ed9\u5355\u96c6\u8fdb\u7a0b\u3002\n\n\u4e24\u4e2a\u4e91\u7aef\u540e\u7aef\u90fd\u662f\u540c\u6b65\u63a5\u53e3\u3002DashScope \u9650\u5236\u4e3a**\u7f16\u7801\u540e\u4e0d\u8d85\u8fc7 10MB\u3001\u65f6\u957f\u4e0d\u8d85\u8fc7 5 \u5206\u949f**\uff08\u5ba2\u6237\u7aef\u5728\u539f\u59cb\u6587\u4ef6\u8d85\u8fc7 7 MiB \u65f6\u62d2\u7edd\u5e76\u63d0\u793a\u6539\u7528\u672c\u5730 SenseVoice\uff09\u3002Groq \u514d\u8d39\u5c42\u5355\u6587\u4ef6\u4e0a\u9650 25MB\uff0c**\u8fbe\u5230\u4e0a\u9650\u7684\u957f\u97f3\u9891\u4f1a\u81ea\u52a8\u5206\u7247**\uff1affmpeg \u5207\u6210 20 \u5206\u949f\u4e00\u6bb5\uff08\u7ea6 9.6MB/\u6bb5\uff09\uff0c\u9010\u6bb5\u8f6c\u5f55\u540e\u6309\u987a\u5e8f\u62fc\u63a5\uff0c\u65f6\u95f4\u6233\u81ea\u52a8\u7d2f\u52a0\u504f\u79fb\uff1b\u5206\u7247\u8fdb\u5ea6\u9010\u6bb5\u843d\u76d8\uff0c\u4e2d\u65ad\u540e\u91cd\u8dd1\u540c\u4e00\u547d\u4ee4\u65ad\u70b9\u7eed\u4f20\uff0c\u4e0d\u91cd\u590d\u63d0\u4ea4\u5df2\u5b8c\u6210\u5206\u7247\uff1b\u9047 429 \u6309 Retry-After/\u6307\u6570\u9000\u907f\u7b49\u5f85\u3002Groq \u8fd4\u56de\u7684\u5206\u6bb5\u65f6\u95f4\u6233\u4f1a\u4f5c\u4e3a\u9644\u5f55\u4fdd\u7559\u3002\u5206\u7247\u548c\u5bb9\u5668\u8f6c\u6362\u9700\u8981\u672c\u673a `ffmpeg`\u3002\n\n## \u4e91\u7aef\u4efb\u52a1\u6062\u590d\n\n\u9ed8\u8ba4\u72b6\u6001\u76ee\u5f55\u4e3a `~/.local/state/chubbyskills/podcast`\uff1b\u8bbe\u7f6e\u4e86 `XDG_STATE_HOME` \u65f6\u4f7f\u7528\u5176\u4e2d\u7684 `chubbyskills/podcast`\u3002\u8fd8\u53ef\u7528 `CHUBBY_PODCAST_STATE_DIR` \u6216 `--state-dir` \u6307\u5b9a\u3002\u76ee\u5f55\u5305\u542b\u4efb\u52a1\u4e0e\u8f6c\u5f55\u5185\u5bb9\uff0c\u5e94\u6309\u97f3\u9891\u5185\u5bb9\u7684\u9690\u79c1\u8981\u6c42\u4fdd\u5b58\u3002\n\n\u76f8\u540c\u97f3\u9891\u3001provider\u3001\u6a21\u578b\u3001\u8bed\u8a00\u548c\u670d\u52a1\u5730\u5740\u518d\u6b21\u8fd0\u884c\u65f6\uff0c\u5df2\u6709\u4efb\u52a1\u7ee7\u7eed\u67e5\u8be2\uff1b\u5df2\u5b8c\u6210\u7ed3\u679c\u53ef\u4ee5\u91cd\u65b0\u5bfc\u51fa\u3002\u63d0\u4ea4\u7ed3\u679c\u4e0d\u660e\u786e\u6216\u670d\u52a1\u7aef\u62a5\u544a\u5931\u8d25\u65f6\uff0c\u4e0d\u81ea\u52a8\u91cd\u65b0\u63d0\u4ea4\u3002\u666e\u901a\u7f51\u7edc\u91cd\u8bd5\u4fdd\u7559\u72b6\u6001\u76ee\u5f55\u5e76\u91cd\u8dd1\u539f\u547d\u4ee4\u5373\u53ef\u3002\n\n`--resubmit` \u660e\u786e\u521b\u5efa\u65b0\u4efb\u52a1\uff0c\u53ef\u80fd\u91cd\u590d\u8ba1\u8d39\u3002\u4e0d\u8981\u7528\u5b83\u89e3\u51b3\u5355\u7eaf\u7684\u8f6e\u8be2\u8d85\u65f6\u3002\u5220\u9664\u72b6\u6001\u76ee\u5f55\u6216\u6539\u53d8\u8f93\u5165\u914d\u7f6e\u4e5f\u53ef\u80fd\u5931\u53bb\u590d\u7528\u6761\u4ef6\u3002\u5ba2\u6237\u7aef\u8d85\u65f6\u4e0d\u7b49\u4e8e\u670d\u52a1\u7aef\u53d6\u6d88\uff0c\u4e5f\u4e0d\u4ee3\u8868\u6ca1\u6709\u8ba1\u8d39\u3002\n\n\u5b8c\u6574\u4ed3\u5e93\u4f7f\u7528\u8bf4\u660e\u3001\u7edf\u4e00\u5165\u5e93\u548c\u9a8c\u8bc1\u8303\u56f4\u89c1[\u4e91\u7aef\u8f6c\u5f55\u8bf4\u660e](https://github.com/chubbyguan/chubbyskills/blob/main/docs/cloud-transcription.md)\u3002\n\n## \u4ea7\u7269\u4e0e\u9650\u5236\n\n\u4ea7\u7269\u4e3a\u5e26 frontmatter\u3001\u6765\u6e90\u548c\u8f6c\u5f55\u540e\u7aef\u6807\u8bb0\u7684 Markdown\u3002\u540e\u7aef\u8fd4\u56de\u53ef\u7528\u5206\u6bb5\u65f6\u4fdd\u7559\u65f6\u95f4\u6233\uff1b\u6ca1\u6709\u65f6\u95f4\u4fe1\u606f\u65f6\u4e0d\u7f16\u9020\u65f6\u95f4\u8f74\u3002\n\n- \u672c\u5730 SenseVoice-Small \u53ea\u8f93\u51fa\u7eaf\u6587\u672c\u5168\u6587\uff0c\u4e0d\u751f\u6210\u9010\u6bb5\u65f6\u95f4\u6233\uff08\u6bb5\u6570\u8bb0\u4e3a 1\uff09\u3002\n- \u672c\u5730 CPU \u63a8\u7406\u8017\u65f6\u53d7\u97f3\u9891\u957f\u5ea6\u548c\u673a\u5668\u914d\u7f6e\u5f71\u54cd\u3002\n- \u8f6c\u5f55\u53ef\u80fd\u6709\u4e13\u6709\u540d\u8bcd\u3001\u6570\u5b57\u6216\u65ad\u53e5\u9519\u8bef\uff0c\u5f15\u7528\u524d\u6838\u5bf9\u539f\u59cb\u97f3\u9891\u3002\n- \u4e91\u7aef\u771f\u5b9e\u53ef\u7528\u6027\u3001\u8d26\u53f7\u6743\u9650\u3001\u97f3\u9891\u517c\u5bb9\u6027\u548c\u8d39\u7528\u5c1a\u9700\u72ec\u7acb\u9a8c\u6536\u3002\n- \u672c skill \u4e0d\u63d0\u4f9b\u901a\u7528\u8bf4\u8bdd\u4eba\u5206\u79bb\u4fdd\u8bc1\u3002\n\n## \u8d21\u732e\u4e0e\u53c2\u8003\n\n\u53ef\u9009\u4e91\u7aef\u8f6c\u5f55\u9700\u6c42\u6700\u521d\u6765\u81ea [binyangzhu000-sudo \u7684 PR #3](https://github.com/chubbyguan/chubbyskills/pull/3) \u548c [Anil-matcha \u7684 PR #5](https://github.com/chubbyguan/chubbyskills/pull/5)\uff08Atlas / MuAPI \u5b9e\u9a8c\u540e\u7aef\uff0c\u73b0\u5df2\u88ab DashScope \u540e\u7aef\u53d6\u4ee3\uff0c\u5f52\u5c5e\u4fdd\u7559\uff09\u3002\n\n- [FunAudioLLM/SenseVoice](https://github.com/FunAudioLLM/SenseVoice)\n- [\u963f\u91cc\u4e91\u767e\u70bc Qwen-ASR API \u53c2\u8003](https://help.aliyun.com/zh/model-studio/qwen-asr-api-reference)\n- [Groq Speech to Text \u6587\u6863](https://console.groq.com/docs/speech-to-text)\n\n## \u5408\u89c4\u58f0\u660e\n\n\u8bf7\u9075\u5b88\u6765\u6e90\u5e73\u53f0\u6761\u6b3e\u5e76\u5c0a\u91cd\u5185\u5bb9\u7248\u6743\uff0c\u63a7\u5236\u8bf7\u6c42\u9891\u7387\u3002\u4e91\u7aef\u5904\u7406\u524d\u786e\u8ba4\u81ea\u5df1\u6709\u6743\u5411\u6240\u9009\u670d\u52a1\u63d0\u4ea4\u97f3\u9891\uff1b\u4e0b\u8f7d\u548c\u8f6c\u5f55\u4e0d\u4f1a\u6539\u53d8\u539f\u5185\u5bb9\u7684\u7248\u6743\u5f52\u5c5e\u3002\n","usage":"\u8bfb\u53d6 SKILL.md\uff0c\u5e76\u6309\u9700\u8bfb\u53d6\u914d\u5957\u6587\u4ef6\uff1b\u4e0b\u8f7d ZIP \u83b7\u53d6\u5b8c\u6574\u5305\u3002\u6280\u80fd\u662f\u4efb\u52a1\u65b9\u6cd5\uff0c\u4e0d\u6388\u4e88\u65b0\u589e\u8d26\u6237\u3001\u4ed8\u8d39\u6216\u7cfb\u7edf\u6743\u9650\u3002"}
