# PyChain: A Fully Parallelized PyTorch Implementation of LF-MMI for End-to-End ASR

arXiv: https://arxiv.org/abs/2005.09824 · HTML: https://mlsanity.com/paper/2005.09824/

## Key facts

- Cited by 16 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 20 May 2020.

## Built on by (most cited first, 16 of 16)

- [CTC Variations Through New WFST Topologies](https://mlsanity.com/paper/2110.03098/) (2021)
- [Recent Progress in the CUHK Dysarthric Speech Recognition System](https://mlsanity.com/paper/2201.05845/) (2022)
- [Semantic-WER: A Unified Metric for the Evaluation of ASR Transcript for End Usability](https://mlsanity.com/paper/2106.02016/) (2021)
- [Wake Word Detection with Streaming Transformers](https://mlsanity.com/paper/2102.04488/) (2021)
- [Differentiable Allophone Graphs for Language-Universal Speech Recognition](https://mlsanity.com/paper/2107.11628/) (2021)
- [On lattice-free boosted MMI training of HMM and CTC-based full-context ASR models](https://mlsanity.com/paper/2107.04154/) (2021)
- [Multi-Channel Multi-Speaker ASR Using 3D Spatial Feature](https://mlsanity.com/paper/2111.11023/) (2021)
- [Comparing CTC and LFMMI for out-of-domain adaptation of wav2vec 2.0 acoustic model](https://mlsanity.com/paper/2104.02558/) (2021)
- [Integrating Lattice-Free MMI into End-to-End Speech Recognition](https://mlsanity.com/paper/2203.15614/) (2022)
- [Injecting Text and Cross-lingual Supervision in Few-shot Learning from Self-Supervised Models](https://mlsanity.com/paper/2110.04863/) (2021)
- [GPU-Accelerated Forward-Backward algorithm with Application to Lattice-Free MMI](https://mlsanity.com/paper/2112.00709/) (2021)
- [Effectiveness of Text, Acoustic, and Lattice-based representations in Spoken Language Understanding tasks](https://mlsanity.com/paper/2212.08489/) (2022)
- [Finnish Parliament ASR corpus - Analysis, benchmarks and statistics](https://mlsanity.com/paper/2203.14876/) (2022)
- [Incorporating L2 Phonemes Using Articulatory Features for Robust Speech Recognition](https://mlsanity.com/paper/2306.02534/) (2023)
- [Beyond Universal Transformer: block reusing with adaptor in Transformer for automatic speech recognition](https://mlsanity.com/paper/2303.13072/) (2023)
- [RIR-SF: Room Impulse Response Based Spatial Feature for Target Speech Recognition in Multi-Channel Multi-Speaker Scenarios](https://mlsanity.com/paper/2311.00146/) (2023)

## Abstract (arXiv)

We present PyChain, a fully parallelized PyTorch implementation of end-to-end lattice-free maximum mutual information (LF-MMI) training for the so-called chain models in the Kaldi automatic speech recognition (ASR) toolkit. Unlike other PyTorch and Kaldi based ASR toolkits, PyChain is designed to be as flexible and light-weight as possible so that it can be easily plugged into new ASR projects, or other existing PyTorch-based ASR tools, as exemplified respectively by a new project PyChain-example, and Espresso, an existing end-to-end ASR toolkit. PyChain's efficiency and flexibility is demonstrated through such novel features as full GPU training on numerator/denominator graphs, and support for unequal length sequences. Experiments on the WSJ dataset show that with simple neural networks and commonly used machine learning techniques, PyChain can achieve competitive results that are comparable to Kaldi and better than other end-to-end ASR systems.