#5151st of 51 in VoiceFor voice, speech and music
asrt_speechrecognition
Mandarin speech recognition on TensorFlow Keras, with HTTP and gRPC servers
Live demo ↗
(opens in a new tab)Documentation ↗
(opens in a new tab)Website ↗
(opens in a new tab)Repository on GitHub ↗
(opens in a new tab)
- Stars
- 8.4k
- License
- GPL-3.0
- Last commit
- Apr 2026
- Last release
- May 2022
- Language
- Python
8,305 on 1 Apr 20268,394 on 21 Oct 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
ASRT is a Chinese speech recognition system. A DCNN+CTC acoustic model maps audio of up to 16 seconds to pinyin, and a maximum-entropy hidden Markov language model converts pinyin to Chinese characters. The repo includes training and evaluation scripts, HTTP and gRPC API servers, and client SDKs for several languages.
Who it is for: Developers needing self-hosted Mandarin speech-to-text or their own model
Strengths
- Full pipeline in one repo: training, evaluation, single-file prediction and API servers
- Serves recognition over HTTP and gRPC, with client SDKs for Python, Go, Java and Windows
- Docker image runs CPU-only inference, exposing ports 20001 and 20002
- Default config covers six public Chinese datasets, including THCHS30, AIShell-1 and MagicData
Weaknesses
- Chinese only; the README mentions no other languages
- README reports about 85% pinyin accuracy on the test set for the top model
- Training needs an NVIDIA GPU with 11 GB+ VRAM, 16 GB RAM and 500 GB disk
- Last release was 2022-05; README and docs are mainly in Chinese
What it needs
- RAM ≥ 16 GB
- GPU optional
- Docker
- Needs TensorFlow 2.5+, Python 3.9+, Flask, Waitress, grpcio, CUDA 11.2 and cuDNN 8.1 for GPU training
- Models: DCNN+CTC acoustic models 24, 25, 251, 251bn, pinyin-to-text statistical language model
- port 20001
Also in Voice
See all 51| Rank | Project | Score |
|---|---|---|
| 1 | voicestudioDesktop app for voice cloning, dubbing and dictation with local speech models | 83 out of 100 |
| 2 | xiaozhi-esp32ESP32 firmware for a voice chatbot with MCP device and cloud tool control | 80 out of 100 |
| 3 | whisper.cppC/C++ port of OpenAI Whisper for local speech-to-text transcription | 78 out of 100 |
| #4 | VoiceboxLocal voice studio for cloning, TTS, dictation and agent speech | 71 out of 100 |
| #5 | buzzOffline desktop app for transcribing and translating audio with Whisper | 71 out of 100 |