Z-yq/TensorflowASR
A Tensorflow 2 end-to-end speech recognition system implementing Conformer and CTC models with CPU-optimized inference achieving real-time factor under 0.1.
Not currently ranked — collecting fresh signals.
star history
This repository provides a Tensorflow-based implementation of automatic speech recognition models, primarily using Conformer architecture with CTC decoding. It supports both streaming (online) and offline recognition modes, with features including VAD, denoising, and punctuation restoration. The project focuses on CPU optimization to achieve near real-time performance while maintaining competitive accuracy on benchmarks like Aishell-1.
Frequently asked
- What is Z-yq/TensorflowASR?
- A Tensorflow 2 end-to-end speech recognition system implementing Conformer and CTC models with CPU-optimized inference achieving real-time factor under 0.1.
- Is TensorflowASR open source?
- Yes — Z-yq/TensorflowASR is open source, released under the Apache-2.0 license.
- What language is TensorflowASR written in?
- Z-yq/TensorflowASR is primarily written in Python.
- How popular is TensorflowASR?
- Z-yq/TensorflowASR has 474 stars on GitHub.
- Where can I find TensorflowASR?
- Z-yq/TensorflowASR is on GitHub at https://github.com/Z-yq/TensorflowASR.