PaddleSpeech and RapidASR
RapidASR is a lightweight inference wrapper built on top of FunASR models, making it a complement that simplifies deployment of PaddleSpeech's ASR capabilities across platforms via ONNX Runtime rather than a competitor.
About PaddleSpeech
PaddlePaddle/PaddleSpeech
Easy-to-use Speech Toolkit including Self-Supervised Learning model, SOTA/Streaming ASR with punctuation, Streaming TTS with text frontend, Speaker Verification System, End-to-End Speech Translation and Keyword Spotting. Won NAACL2022 Best Demo Award.
Built on the PaddlePaddle framework, the toolkit implements streaming ASR/TTS systems with rule-based Chinese text normalization, polyphone handling, and tone sandhi processing through a dedicated linguistic frontend. It provides production-ready deployment via CLI, REST API server, and WebSocket streaming server interfaces, with pre-trained models optimized for both accuracy and inference speed across multiple languages including English, Mandarin, and Cantonese.
About RapidASR
RapidAI/RapidASR
📣 商用级开源语音自动识别程序库，开箱即用，全平台支持，中英文混合识别。A Cross-platform implementation of ASR inference. It's based on ONNXRuntime and FunASR. We provide a set of easier APIs to call ASR models.
Leverages Alibaba's Paraformer model with ONNX Runtime inference for optimized performance, supporting batch processing and multiple input formats (file paths, numpy arrays, audio lists). Integrates with a complementary pipeline including RapidVad for voice activity detection and RapidPunc for punctuation restoration. Offers both Python (3.6+) and C++ implementations across Linux, Windows, and macOS.
Scores updated daily from GitHub, PyPI, and npm data. How scores work