[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"news-slug-granite-speech-5-collection-update":3,"topics-all":38,"news-related-35e9f682-7952-4f6a-a8e9-7f8bdd47e791":57},{"id":4,"title":5,"summary":6,"content":7,"original_url":8,"source_id":9,"tags":10,"translations":24,"news_slug":31,"published_at":32,"created_at":33,"modified_at":34,"is_published":35,"publish_type":36,"image_url":14,"view_count":37},"35e9f682-7952-4f6a-a8e9-7f8bdd47e791","IBM 语音三件套更新:2B 旗舰 ASR 精度开源第二","IBM 10 月 21 日更新 Granite Speech 语音三件套:新 2B 旗舰覆盖六语转写互译,39 万小时音频训练,平均英语 WER 5.06%;470M 边缘模型扩到七语、提速 1.6 倍;2B-plus 说话人错误率降 7%、时间戳偏移减半,全部 Apache 2.0 开源。","10 月 21 日,IBM 在 Hugging Face 上一次更新了 Granite Speech 家族的三件套:新旗舰 granite-speech-5.0-2b,加上两个存量模型的修订版——边缘端的 granite-speech-5.0-470m-turboctc 和会议场景的 granite-speech-4.1-2b-plus。一次动三个模型,在语音赛道并不多见,摆明了是要把企业语音的三类需求一次配齐。\n\n## 三件套各自管什么\n\n**2B 旗舰管精度。** granite-speech-5.0-2b 是 4.1-2b 的同参数量继任者,支持英语、法语、德语、西班牙语、葡萄牙语、日语六种语言转写,并与英语互译,另加英语到意大利语和普通话的翻译方向。训练数据为 39 万小时公开音频,外加为日语 ASR、关键词偏置和翻译专门构造的合成数据。它支持最多 2000 词的关键词偏置表,把产品名、人名、缩写\"喂\"给模型;标点和大小写可按需请求,官方测试集上大小写 F1 为 89.7 到 99.5。\n\n**470M TurboCTC 管速度。** 这个 8 月首发时还是英语专供的 470M 编码器,这次改版扩到七种语言(英、日、西、德、法、意、葡),转写结果原生带标点和大小写,并针对混响环境强化了训练。提速来自把单个 Conformer 块编译后在全栈复用:Open ASR Leaderboard 上的 RTFx 从发布时的约 12600 提到 19870,约 1.6 倍,精度不变。\n\n**4.1-2b-plus 管\"谁在说话\"。** 架构、提示格式和 API 都没动,是原地换权重的 drop-in 升级。官方数据显示:词级说话人错误率 WDER 从 4.40 降到 4.10,降 7%;说话人归因转写错误率 cpWER 从 30.50 降到 28.48;长语音测试集 WER 从 13.53 降到 10.22,降 24%;20 词偏置表下的关键词错误率从 11.79 降到 6.48,降 45%;全语音平均时间戳偏移从 210.86 砍到 98.54,直接减半——对字幕和检索场景,这是硬收益。\n\n## 数字之外的三点观察\n\n第一,**企业语音栈正在分层**。精度、边缘速度、说话人归属三类需求,IBM 用三个分工明确的模型覆盖,而不是塞进一个\"全能模型\"。对照同周的 Qwen-Audio-3.1-ASR 和 Voxtral 3,开源语音赛道的供给正在快速拥挤。\n\n第二,**许可和推理栈都给足了**。三个模型统一 Apache 2.0、原生支持 Transformers,其中 2B 旗舰还适配 vLLM、llama.cpp 和 Apple Silicon 的 mlx-audio,TurboCTC 另有浏览器 WebGPU 流式 demo。对不想被云 API 锁定的团队,这是一条完整的自托管路径。\n\n第三,**榜单数字要看口径**。\"平均英语 WER 5.06%、位居开源权重模型第二\"是 IBM 官方在 Open ASR Leaderboard 公共英语短测集上的自报成绩,截止 10 月 21 日;1.6 倍提速同样出自官方口径,第三方复测尚未出现。采信数字时保持归因,是读厂商榜单的基本功。\n\n参考:IBM Granite 官方博客 https:\u002F\u002Fhuggingface.co\u002Fblog\u002Fibm-granite\u002Fgranite-speech-5\n\n对企业用户来说,信号很清楚:语音这条线正在重演 LLM 已走过的剧本——开源全家桶、明确分工、边缘可跑。下一个看点,是谁先把\"转写+归属+翻译\"做成真正顺手的端到端产品。","https:\u002F\u002Fhuggingface.co\u002Fblog\u002Fibm-granite\u002Fgranite-speech-5","227ef3c2-ea27-421e-86e8-ee005617a36e",[11,15,18,21],{"id":12,"name":13,"slug":13,"description":14,"color":14},"c4455801-0273-477d-9e92-00b4addfa439","asr",null,{"id":16,"name":17,"slug":17,"description":14,"color":14},"7e89b5cc-57db-4f37-bc6d-28919a73931c","model-release",{"id":19,"name":20,"slug":20,"description":14,"color":14},"499f4b56-819d-49a3-9609-33e775143b86","multimodal",{"id":22,"name":23,"slug":23,"description":14,"color":14},"b9bd9039-fcdb-41a8-b85b-fc1587def2b9","open-source",[25],{"id":26,"lang":27,"title":28,"summary":29,"content":30},"6dfa0d64-3776-4f8c-ac9a-a9ee08407536","en","IBM Updates Granite Speech Trio: 2B Flagship, Faster Edge Model","IBM's Granite Speech refresh: a 390K-hour 2B flagship, a 470M edge model, 7 languages, 1.6x faster, and 2B-plus cutting speaker error 7%. Apache 2.0.","On October 21, IBM updated its Granite Speech family on Hugging Face in one move: a new flagship, granite-speech-5.0-2b, plus revised versions of two models the community already runs — the edge-oriented granite-speech-5.0-470m-turboctc and the meeting-focused granite-speech-4.1-2b-plus. Updating three models at once is unusual in the speech track, and it signals an attempt to cover the three core enterprise speech needs in a single sweep.\n\n## What each model handles\n\n**The 2B flagship handles accuracy.** granite-speech-5.0-2b succeeds granite-speech-4.1-2b at the same parameter count. It transcribes English, French, German, Spanish, Portuguese, and Japanese, translates to and from English in those languages, and adds English-to-Italian and English-to-Mandarin directions. Training used 390,000 hours of public audio plus synthetic data built for Japanese ASR, keyword biasing, and translation. It supports keyword biasing lists of up to 2,000 terms — feeding product names, people, and acronyms directly to the model — and punctuation and capitalization on request, with capitalization F1 ranging from 89.7 to 99.5 on IBM's test sets.\n\n**The 470M TurboCTC handles speed.** Launched in August as an English-only 470M encoder, this revision expands to seven languages (English, Japanese, Spanish, German, French, Italian, Portuguese), outputs native punctuation and capitalization, and adds training that targets reverberant audio. The speedup comes from compiling a single Conformer block and reusing its graph across the stack: RTFx on the Open ASR Leaderboard rose from roughly 12,600 at launch to 19,870 — about 1.6x faster at the same accuracy.\n\n**The 4.1-2b-plus handles \"who spoke.\"** Architecture, prompt format, and API are unchanged, making it a drop-in weight swap. Official numbers: word diarization error rate (WDER) dropped from 4.40 to 4.10, a 7% cut; speaker-attributed WER (cpWER) fell from 30.50 to 28.48; long-form WER dropped 24% (13.53 to 10.22); biased-keyword errors fell 45% with 20-term lists (11.79 to 6.48); and average timestamp shift on all speech was cut from 210.86 to 98.54 — more than half, a direct win for captions and search.\n\n## Three observations beyond the numbers\n\nFirst, **the enterprise speech stack is stratifying**. Accuracy, edge speed, and speaker attribution are three different needs, and IBM covers them with three purpose-built models rather than one \"do-everything\" model. Set against the same-week Qwen-Audio-3.1-ASR and Voxtral 3 releases, open speech supply is getting crowded fast.\n\nSecond, **licensing and inference stacks are fully covered**. All three ship under Apache 2.0 with native Transformers support; the 2B flagship also runs on vLLM, llama.cpp, and mlx-audio for Apple Silicon, and TurboCTC has a browser WebGPU streaming demo. For teams avoiding cloud API lock-in, this is a complete self-hosted path.\n\nThird, **read the leaderboard numbers with attribution in mind**. The \"5.06% average English WER, No.2 among open-weight models\" figure is IBM's own reporting on the Open ASR Leaderboard's public English short-form sets as of October 21; the 1.6x speedup is likewise an official claim, and independent retests have not yet appeared.\n\nReference: IBM Granite official blog https:\u002F\u002Fhuggingface.co\u002Fblog\u002Fibm-granite\u002Fgranite-speech-5\n\nFor enterprise users the signal is clear: speech is replaying the script LLMs already ran — open-weight family releases, clear division of labor, edge-deployable. The next question is who first turns \"transcribe + attribute + translate\" into a genuinely smooth end-to-end product.","granite-speech-5-collection-update","2026-10-22T21:08:48Z","2026-10-22T21:10:13.860911Z","2026-10-22T21:10:13.860922Z",true,"agent",10,[39,48],{"slug":40,"tag_slug":40,"title_zh":41,"title_en":42,"intro_zh":43,"intro_en":44,"id":45,"is_active":35,"created_at":46,"modified_at":47},"ai-for-science","AI for Science 2026：从 UniPert 到 GPT-Rosalind 的硬核进化","AI for Science 2026: from UniPert to GPT-Rosalind","生命科学、化学材料、物理世界模型——AI 正在从\"语言工具\"变成\"实验伙伴\"。本专题收录 AI 在三大科学方向的关键节点：UniPert 统一基因与化学扰动空间、GPT-Rosalind 端到端生命科学推理、达摩院 AI 智能体 28 小时找到 4 种超导新材料、Anthropic Claude Science 把工作台做成标准品。","From language tool to lab partner — AI is reshaping life sciences, chemistry\u002Fmaterials, and physical world models. This topic covers the key milestones: UniPert unifying genetic-chemical perturbation spaces, GPT-Rosalind's end-to-end life-sciences reasoning, DAMO's AI agent discovering 4 superconducting materials in 28 hours, and Anthropic's Claude Science workbench going mainstream.","988a4300-5fab-41c4-b5d8-63711a2dc757","2026-09-10T01:34:15.296649Z","2026-09-10T01:34:15.296663Z",{"slug":49,"tag_slug":49,"title_zh":50,"title_en":51,"intro_zh":52,"intro_en":53,"id":54,"is_active":35,"created_at":55,"modified_at":56},"h3-series","MiniMax H3 系列：从开源权重到 35 倍吞吐","MiniMax H3 Series: from open weights to 35x throughput","MiniMax H3 自 2026 年 8 月开源以来节奏密集：官方把生成、参考与编辑收回一个模型；ComfyUI 当天压进 RTX 3060；摩尔线程 3 小时完成国产 GPU 适配；fal 后训练版把吞吐拉到 35 倍；FastH3 蒸馏再砍推理成本。本专题持续追踪 H3 的发布—开源—蒸馏—部署全链路。","Since MiniMax open-sourced H3 in August 2026 the pace has been relentless: one unified omni-modal model, same-day ComfyUI support down to an RTX 3060, a 3-hour Day-0 port to Moore Threads GPUs, fal's post-trained H3 Max at 35x throughput, and FastH3 distillation cutting inference cost further. This topic tracks the full H3 chain — release, open weights, distillation, deployment.","83ef0daa-3c31-4cb3-86ed-e5ee58654d5f","2026-09-08T07:33:19.942193Z","2026-09-08T07:33:19.942209Z",{"items":58},[59,64,69,74,79,84],{"id":60,"title":61,"news_slug":62,"published_at":63},"d83c4910-6dce-4171-8a44-9f68c9f841af","Qwen-Audio-3.1-ASR开源:转写自带说话人","qwen-audio-3-1-asr-speaker-diarization","2026-10-22T17:13:29+00:00",{"id":65,"title":66,"news_slug":67,"published_at":68},"bbbda7ff-b1c7-4142-b37f-2b3fd0547e02","Fibo 2.0 开源:生成编辑合一,登顶开源编辑榜","bria-fibo-2-0-unified-editing-generation","2026-10-21T21:08:11+00:00",{"id":70,"title":71,"news_slug":72,"published_at":73},"bef7a773-36fb-406b-949d-8d728b97b242","Matrix-Game 4.0 开源:单卡 60FPS 的实时可玩游戏世界","matrix-game-4-0-open-source-60fps","2026-10-21T13:09:32+00:00",{"id":75,"title":76,"news_slug":77,"published_at":78},"b44fb160-9027-44cd-b355-a66869004abc","Nemotron 3.5 Super VL 开源:120B 总参只激活 12B","nemotron-3-5-super-vl-open-120b-moe","2026-10-20T17:11:56+00:00",{"id":80,"title":81,"news_slug":82,"published_at":83},"faf5c85e-ed83-432a-8b55-80c9356862c5","腾讯优图开源5B全模态解析:文档榜96.96超GPT-5.2","youtu-parsing-omni-5b","2026-10-19T19:09:27+00:00",{"id":85,"title":86,"news_slug":87,"published_at":88},"405b8f47-8862-4684-ab0e-d552ce093072","MJev-Spatial:空间理解不生成文字,单次前向快5.6倍","mjev-spatial-classification-spatial-intelligence","2026-10-18T17:14:25+00:00"]