thu-spmi/CAT has 369 GitHub stars and 79 forks. GitHot measured 0 stars in 24 hours. Last checked 2026-11-18 00:41 UTC. Last push 2026-02-05.
What you can do
- CAT contains a full-fledged CUDA/C/C++ implementation of CTC-CRF loss function binding to PyTorch.
- One-stop CTC/CTC-CRF/RNN-T/LM training & inference. See the templates.
- Flexible configuration with JSON. Check the guideline for configuration.
- Scalable and extensible. It is easy to be extended to train tens of thousands of speech data and add new models and tasks.
Measured star growth
- in 24h
- 0
- in 7 days
- Collecting history
- in 15 days
- Collecting history
- in 1 month
- Collecting history
Net star changes over overlapping periods, not daily totals. Missing measurements are left blank. Stars measure attention, not software quality.
About
CAT is more than a CRF-based ASR toolkit: it provides a complete workflow for data-efficient end-to-end ASR, supporting CTC, CTC-CRF, RNN-T, and language-model training and inference.
ProductivityREADME.md
Original
CAT: CRF-based ASR Toolkit
CAT provides a complete workflow for CRF-based data-efficient end-to-end speech recognition.
See What's New for recently added functionalities and features!
Overview
CAT aims at combining the advantages of both the hybrid and the E2E ASR approches to achieve data-efficiency, by judiciously examining the pros and cons of modularity versus unified neural network, separate optimization versus joint optimization. CAT advocates global normalization modeling and discriminative training in the framework of Conditional Random Field (CRF), currently with Connectionist Temporal Classification (CTC) inspired state topology.
Features
CAT contains a full-fledged CUDA/C/C++ implementation of CTC-CRF loss function binding to PyTorch.
One-stop CTC/CTC-CRF/RNN-T/LM training & inference. See the templates.
Flexible configuration with JSON. Check the guideline for configuration.
Scalable and extensible. It is easy to be extended to train tens of thousands of speech data and add new models and tasks.
Installation
Dependencies
CUDA compatible device, NVIDIA driver installed and CUDA lib available.
PyTorch:
>=1.9.0is required. Installation guide from PyTorchKaldi [optional, but recommended]: used for speech data preparation and some FST-related operations. This is optional for most of the basic functions. Required only when you want to conduct CTC-CRF training.
Besides Kaldi, you could use
torchaudiofor feature extraction. Take a look at data.sh for how to prepare data withtorchaudio.
Clone and install CAT
git clone https://github.com/thu-spmi/CAT.git && cd CAT # Get installation helping message ./install.sh -h # Install with default configurations #./install.sh
Getting started
To get started with this project, please refer to TEMPLATE for tutorial.
ASR results
| dataset | evaluation sets | performance |
|---|---|---|
| AISHELL-1 | dev / test | 3.93 / 4.22 |
| Commonvoice German | test | 9.8 |
| Librispeech | test-clean / test-other | 1.94 / 4.39 |
| Switchboard | switchboard / callhome | 6.9 / 14.5 |
| THCHS30 | test | 6.01 |
| Wenetspeech | test-net / test-meeting | 9.32 / 14.66 |
| WSJ | eval92 / dev93 | 2.77 / 5.68 |
Further reading
- Some tips about the usage of third party tools
- Tutorial on building your first CAT project (yesno)
- Step-by-step workflow for CAT-v2
Citation
@inproceedings{xiang2019crf,
title={CRF-based single-stage acoustic modeling with CTC topology},
author={Xiang, Hongyu and Ou, Zhijian},
booktitle={IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)},
pages={5676--5680},
year={2019},
organization={IEEE}
}
@inproceedings{an2020cat,
title={CAT: A CTC-CRF based ASR toolkit bridging the hybrid and the end-to-end approaches towards data efficiency and low latency},
author={An, Keyu and Xiang, Hongyu and Ou, Zhijian},
booktitle={INTERSPEECH},
pages={566--570},
year={2020}
}