Many machine learning tasks can be expressed as the transformation---or \emph{transduction}---of input sequences into output sequences: speech recognition, machine translation, protein secondary structure prediction and text-to-speech to name but a few. One of the key challenges in sequence transduction is learning to represent both the input and output sequences in a way that is invariant to sequential distortions such as shrinking, stretching and translating. Recurrent neural networks (RNNs) are a powerful sequence learning architecture that has proven capable of learning such representations. However RNNs traditionally require a pre-defined alignment between the input and output sequences to perform transduction. This is a severe limitation since \emph{finding} the alignment is the most difficult aspect of many sequence transduction problems. Indeed, even determining the length of the output sequence is often challenging. This paper introduces an end-to-end, probabilistic sequence transduction system, based entirely on RNNs, that is in principle able to transform any input sequence into any finite, discrete output sequence. Experimental results for phoneme recognition are provided on the TIMIT speech corpus.
Install required libraries:
bash install.sh
download LibriSpeech http://www.openslr.org/12
bash scripts/download_librispeech.sh ${DATA_ROOT_DIR}
preprocess LibriSpeech
bash scripts/preprocess_librispeech.sh ${DATA_ROOT_DIR}
sed -i "s#MODIFY_DATASET_DIR#${DATA_ROOT_DIR}/LibriSpeech#g" configs/baseline_v3-1023sp.yaml
mkdir -p output/
bash scripts/train_rnnt_1x8.sh output/ ${DATA_ROOT_DIR}/LibriSpeech
Following conditions were tested, you can run any of them below:
FP32 | |
---|---|
single card | train_rnnt_1x1.sh |
4 cards | train_rnnt_1x4.sh |
8 cards | train_rnnt_1x8.sh |
GPUs | FP16 | FPS | WER |
---|---|---|---|
1x8 | False | 20 | 0.058 |
https://github.com/mlcommons/training/tree/master/rnn_speech_recognition/pytorch
此处可能存在不合适展示的内容,页面不予展示。您可通过相关编辑功能自查并修改。
如您确认内容无涉及 不当用语 / 纯广告导流 / 暴力 / 低俗色情 / 侵权 / 盗版 / 虚假 / 无价值内容或违法国家有关法律法规的内容,可点击提交进行申诉,我们将尽快为您处理。