You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
PaddleSpeech/examples/tiny
Hui Zhang 258307df9b
fix egs bugs (#552)
4 years ago
..
conf Support paddle 2.x (#538) 4 years ago
local fix egs bugs (#552) 4 years ago
.gitignore Support paddle 2.x (#538) 4 years ago
README.md Refactor CTC module, add embedding and fix log (#549) 4 years ago
path.sh Support paddle 2.x (#538) 4 years ago
run.sh Support paddle 2.x (#538) 4 years ago

README.md

Tiny Example

  1. source path.sh
  2. set CUDA_VISIBLE_DEVICES as you need.
  3. demo scrpt is bash run.sh. You can run commond separately as needed.

Steps

  • Prepare the data

    bash local/data.sh
    

    data.sh will download dataset, generate manifests, collect normalizer's statistics and build vocabulary. Once the data preparation is done, you will find the data (only part of LibriSpeech) downloaded in ${MAIN_ROOT}/dataset/librispeech and the corresponding manifest files generated in ${PWD}/data as well as a mean stddev file and a vocabulary file. It has to be run for the very first time you run this dataset and is reusable for all further experiments.

  • Train your own ASR model

    bash local/train.sh
    

    train.sh will start a training job, with training logs printed to stdout and model checkpoint of every pass/epoch saved to ${PWD}/checkpoints. These checkpoints could be used for training resuming, inference, evaluation and deployment.

  • Case inference with an existing model

    bash local/infer.sh
    

    infer.sh will show us some speech-to-text decoding results for several (default: 10) samples with the trained model. The performance might not be good now as the current model is only trained with a toy subset of LibriSpeech. To see the results with a better model, you can download a well-trained (trained for several days, with the complete LibriSpeech) model and do the inference.

  • Evaluate an existing model

    bash local/test.sh
    

    test.sh will evaluate the model with Word Error Rate (or Character Error Rate) measurement. Similarly, you can also download a well-trained model and test its performance:

  • Export jit model

    bash local/export.sh ckpt_path saved_jit_model_path
    
  • Tune hyper paerameter

    bash local/tune.sh