Tencent’s Hy-SimulMT translates Chinese and English while people are talking. One catch: speech output needs a configured TTS service.
Tencent-Hunyuan/Hy-SimulMT
Would you trust it in a live meeting without checking the translation? @grok t.co
3 likes · 52 views
10/26 Hy-SimulMT公開。中国語↔英語を、話し終わる前から順次翻訳する30B／稼働3Bモデル。会議・講演の同時通訳UIの研究向けです。原文も返すInterleave版や量子化版を公開。Apache 2.0、第三者部品は別条件。音声出力には別途TTS設定が必要です。
Tencent-Hunyuan/Hy-SimulMT
#翻訳AI #AI開発
1 like · 13 views
牛皮！
腾讯开源了一个同声传译模型
地址：Tencent-Hunyuan/Hy-SimulMT
中英互译、音色克隆、3秒延时
“Hy-SimulMT 是腾讯混元推出的实时同声传译大模型，支持中文与英文之间的双向语音到语音翻译。说话人的音频可以持续输入，模型边听边译，同步给出译文文字与语音，并支持实时音色克隆，在准确传达信息的同时保留说话人的声音特点。
面向会议、演讲与访谈等实时交流场景，Hy-SimulMT 希望在准确传达信息的同时，减少等待对交流节奏的影响。围绕这一目标，我们结合公开基准测试、真实场景音频和专业同传译员对照，评估模型在翻译质量与响应速度之间的平衡。
在本次 RealSI 长音频评测中，Hy-SimulMT 的 StreamLAAL 约为 3 秒，为图示参评系统中的最低值；GEMBA 得分高于所对比的开源模型，并超过五个商业 API 中的四个。”
3 likes · 155 views
腾讯混元开源了个同传，10 月 22 日挂的仓库，延迟 3 秒，人做同传得 4 到 5 秒，分数还压过五个商业系统!!
结果五天过去，就几颗星。
这俩对不上啊朋友们!!!
Tencent-Hunyuan/Hy-SimulMT
5 likes · 400 views
腾讯混元开源 Hy-SimulMT，中英双向流式语音同传：
在 RealSI 长音频上测得约 3 秒的翻译滞后，同时放出了模型权重、流式推理服务。接上 TTS，还能用说话人的音色读出译文。
代码：Tencent-Hunyuan/Hy-SimulMT
5 likes · 2 reposts · 388 views
tencent/Hy-SimulMT-30B-A3B · Hugging Face
huggingface.co
huggingface.co
Hy-SimulMT is a real-time simultaneous translation model developed by Tencent Hunyuan, supporting bidirectional speech-to-speech translation between Chinese and English. It continuously receives incoming audio and translates as the speaker talks, producing both translated text and speech. With real-time voice cloning, it preserves the speaker's vocal characteristics while conveying the message accurately.
Designed for live communication settings such as meetings, speeches, and interviews, Hy-SimulMT aims to convey information accurately while reducing the disruption that waiting introduces into a conversation. To evaluate this balance between translation quality and response speed, we combine public benchmarks, real-world audio, and comparisons with professional simultaneous interpreters.
In our RealSI long-form audio evaluation, Hy-SimulMT achieved a StreamLAAL of approximately 3 seconds, the lowest among the systems shown in the comparison. Its GEMBA score exceeded those of the compared open-source models and four of the five commercial APIs.
The GitHub repository provides the SimulInfer inference service and SimulTrain training tools. Speech output and voice cloning require a configured TTS service. You can also try Hy-SimulMT in the Hy Translate App.
107 points · 10 comments
Mentions
Tencent’s Hy-SimulMT translates Chinese and English while people are talking. One catch: speech output needs a configured TTS service. Tencent-Hunyuan/Hy-SimulMT Would you trust it in a live meeting without checking the translation? @grok t.co
10/26 Hy-SimulMT公開。中国語↔英語を、話し終わる前から順次翻訳する30B／稼働3Bモデル。会議・講演の同時通訳UIの研究向けです。原文も返すInterleave版や量子化版を公開。Apache 2.0、第三者部品は別条件。音声出力には別途TTS設定が必要です。 Tencent-Hunyuan/Hy-SimulMT #翻訳AI #AI開発
牛皮！ 腾讯开源了一个同声传译模型 地址：Tencent-Hunyuan/Hy-SimulMT 中英互译、音色克隆、3秒延时 “Hy-SimulMT 是腾讯混元推出的实时同声传译大模型，支持中文与英文之间的双向语音到语音翻译。说话人的音频可以持续输入，模型边听边译，同步给出译文文字与语音，并支持实时音色克隆，在准确传达信息的同时保留说话人的声音特点。 面向会议、演讲与访谈等实时交流场景，Hy-SimulMT 希望在准确传达信息的同时，减少等待对交流节奏的影响。围绕这一目标，我们结合公开基准测试、真实场景音频和专业同传译员对照，评估模型在翻译质量与响应速度之间的平衡。 在本次 RealSI 长音频评测中，Hy-SimulMT 的 StreamLAAL 约为 3 秒，为图示参评系统中的最低值；GEMBA 得分高于所对比的开源模型，并超过五个商业 API 中的四个。”
腾讯混元开源了个同传，10 月 22 日挂的仓库，延迟 3 秒，人做同传得 4 到 5 秒，分数还压过五个商业系统!! 结果五天过去，就几颗星。 这俩对不上啊朋友们!!! Tencent-Hunyuan/Hy-SimulMT
腾讯混元开源 Hy-SimulMT，中英双向流式语音同传： 在 RealSI 长音频上测得约 3 秒的翻译滞后，同时放出了模型权重、流式推理服务。接上 TTS，还能用说话人的音色读出译文。 代码：Tencent-Hunyuan/Hy-SimulMT
tencent/Hy-SimulMT-30B-A3B · Hugging Face huggingface.co huggingface.co Hy-SimulMT is a real-time simultaneous translation model developed by Tencent Hunyuan, supporting bidirectional speech-to-speech translation between Chinese and English. It continuously receives incoming audio and translates as the speaker talks, producing both translated text and speech. With real-time voice cloning, it preserves the speaker's vocal characteristics while conveying the message accurately. Designed for live communication settings such as meetings, speeches, and interviews, Hy-SimulMT aims to convey information accurately while reducing the disruption that waiting introduces into a conversation. To evaluate this balance between translation quality and response speed, we combine public benchmarks, real-world audio, and comparisons with professional simultaneous interpreters. In our RealSI long-form audio evaluation, Hy-SimulMT achieved a StreamLAAL of approximately 3 seconds, the lowest among the systems shown in the comparison. Its GEMBA score exceeded those of the compared open-source models and four of the five commercial APIs. The GitHub repository provides the SimulInfer inference service and SimulTrain training tools. Speech output and voice cloning require a configured TTS service. You can also try Hy-SimulMT in the Hy Translate App.