Graves 2013 experiments

File description

• model.py: rnnt joint model
• model2012.py: graves2012 model
• train_rnnt.py: rnnt training script
• train_ctc.py: ctc acoustic model training script
• eval.py: rnnt & ctc decode
• DataLoader.py: kaldi feature loader

Run

• Extract feature link kaldi timit example dirs (local steps utils ) excute run.sh to extract 40 dim fbank feature run feature_transform.sh to get 123 dim feature as described in Graves2013

• Train CTC acoustic model

  python train_ctc.py --lr 1e-3 --bi --dropout 0.5 --out exp/ctc_bi_lr1e-3 --schedule

• Train RNNT joint model

  python train_rnnt.py --lr 4e-4 --bi --dropout 0.5 --out exp/rnnt_bi_lr4e-4 --schedule

• Decode

  python eval.py <path to best model> [--ctc] --bi

Results

Requirements

• Python 3.6
• PyTorch >= 0.4
• numpy 1.14
• warp-transducer

Reference

• RNN Transducer (Graves 2012): Sequence Transduction with Recurrent Neural Networks
• RNNT joint (Graves 2013): Speech Recognition with Deep Recurrent Neural Networks
• (PyTorch End-to-End Models for ASR)[https://github.com/awni/speech]
• (A Fast Sequence Transducer GPU Implementation with PyTorch Bindings)[https://github.com/HawkAaron/warp-transducer/tree/add_network_accelerate]