谷歌今天发布了名为Translatotron 3的新AI模型，该模型可以实现语音对语音的同声传译翻译。Translatotron 3是第一个完全无监督直接语音到语音翻译的端对端模型。该模型的研究人員称，它实现了“第一个完全无监督直接语音到语音翻译的端对端模型”。Translatotron 3可以绕过文本转换步骤，实现更高效的语音翻译。
谷歌Translatotron 3模型的推出，标志着AI语音翻译领域的重要突破。该模型的出现，可以让用户更便捷、更高效地进行语音翻译。未来，Translatotron 3模型将在多个领域得到应用，如谷歌翻译、谷歌文档、谷歌语音识别等。
新闻翻译：
Google has just released a new AI model called Translatotron 3, which can achieve simultaneous translation of voice to voice without any textual conversion steps. Translatotron 3 is the first end-to-end model for direct voice to voice translation that is completely unsupervised. The researchers have called it the “first completely unsupervised direct voice to voice translation end-to-end model.” Translatotron 3 can bypass the text transformation process, achieving more efficient voice translation.
Google’s Translatotron 3 model is a significant breakthrough in the field of AI voice translation. This model will make it easier and more efficient for users to perform voice translation. In the future, Translatotron 3 will be applied in many areas, such as Google translation, Google Docs, and Google speech recognition.
【来源】https://www.ithome.com/0/736/587.htm
Views: 1