Tencent Yuanbao Releases Hy ASR 3.0 Preview Speech Recognition Model: Can Understand Various Dialects and Complex Environments
Tencent
Tencent Yuanbao has launched the Hy ASR 3.0 preview speech recognition model, capable of recognizing various dialects and performing robustly in complex acoustic environments. The model is based on a large-scale audio foundation model with a multi-layer streaming architecture, achieving high accuracy and low latency.
Tencent Yuanbao, Tencent's AI assistant, has introduced the Hy ASR 3.0 preview speech recognition model. The model is designed to accurately transcribe speech across various Chinese dialects and in challenging acoustic conditions, such as noisy environments. It leverages a large-scale audio foundation model with a multi-layer streaming architecture, enabling high recognition accuracy and low latency. The preview release indicates Tencent's ongoing efforts to enhance its AI capabilities in speech processing.
Source: Tencent Hunyuan (GNews) —
original