Alexander-H-Liu/End-to-end-ASR-Pytorch

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

1208 stars · 314 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

A PyTorch implementation of end-to-end automatic speech recognition (ASR), formerly known as Listen, Attend and Spell. It supports seq2seq and CTC-based models with techniques like subword encoding, RNN language models, and joint CTC-attention decoding.

Use cases

• train an end-to-end speech recognition model in pytorch
• implement listen attend and spell ASR
• build a CTC-based speech recognizer
• run beam search decoding for ASR
• train an RNN language model for joint ASR decoding
• experiment with subword tokenization for speech recognition

When to choose

• you want a research-oriented PyTorch codebase for end-to-end ASR
• you need seq2seq, CTC, or hybrid CTC-attention speech recognition models
• you want on-the-fly feature extraction with torchaudio and TensorBoard visualization

When to avoid

• you need a production-ready, actively maintained ASR system
• you want pretrained models or a simple inference API without training
• you lack a high-end GPU, which is required for training

Facets

library · maturity maintenance

speech-recognition machine-learning deep-learning speech-processing machine-learning deep-learning python asr pytorch listen-attend-spell ctc seq2seq beam-search gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="Alexander-H-Liu/End-to-end-ASR-Pytorch")