Based on the FunASR open-source toolset, it supports multiple speech recognition models: - sensevoice-small: (default) Fastest speech recognition, supports multilingual mixing and emotion recognition, with moderate accuracy. - paraformer-large: High-precision Chinese speech recognition, built-in VAD + punctuation + timestamp, limited support for other languages, moderate speed. - fun-asr-nano: Supports recognition of Chinese dialects, singing, etc. Optimized for far-field high-noise recognition. Slowest speed, moderate accuracy, sometimes prone to hallucination-like false recognition. For detailed model features, please refer to the official documentation: https://help.aliyun.com/zh/model-studio/recording-file-recognition The AI Pod provides: - WebUI - Compatible with OpenAI's API, allowing direct access to speech recognition services from various clients! - Compatible with Alibaba Cloud Speech Recognition RESTful API For detailed usage instructions, please refer to the application homepage and documentation. Note: The Whisper model built into the AI Pod has higher accuracy and support for multiple languages than these models, but it is not as fast as the first two. In short: For speech recognition, one API service is enough, balancing speed and accuracy!
