This example contains code used to train a [Conformer](http://arxiv.org/abs/2008.03802) model with [Aishell dataset](http://www.openslr.org/resources/33)
This example contains code used to train a Transformer or [Conformer](http://arxiv.org/abs/2008.03802) model with [Aishell dataset](http://www.openslr.org/resources/33)
## Overview
@ -14,8 +14,7 @@ All the scirpts you need are in the ```run.sh```. There are several stages in th
| 3 | Test the final model performance |
| 4 | Get ctc alignment of test data using the final model |
| 5 | Infer the single audio file |
| 51 | (Not supported at now) Transform the dynamic graph model to static graph model |
| 101 | (Need further installation) Train language model and Build TLG |
You can choose to run a range of stages by setting the ```stage``` and ```stop_stage ``` .
@ -24,6 +23,7 @@ For example, if you want to execute the code in stage 2 and stage 3, you can run
```bash
bash run.sh --stage 2 --stop_stage 3
```
Or you can set ```stage``` equal to ```stop-stage``` to only run one stage.
For example, if you only want to run ```stage 0```, you can use the script below:
@ -38,9 +38,11 @@ The document below will describe the scripts in the ```run.sh``` in detail.
## The environment variables
The path.sh contains the environment variable.
```bash
source path.sh
```
This script needs to be run firstly.
And another script is also needed:
@ -65,11 +67,11 @@ Some local variables are set in the ```run.sh```.
```avg_num``` denotes the number K of top-K models you want to average to get the final model.
```ckpt``` denotes the checkpoint prefix of the model, e.g. "conformer"
```audio file``` denotes the file path of the single file you want to infer in stage 6
You can set the local variables when you use the ```run.sh```
```ckpt``` denotes the checkpoint prefix of the model, e.g. "conformer"
You can set the local variables (except ```ckpt```) when you use the ```run.sh```
For example, you can set the ```gpus``` and ``avg_num`` when you use the command line.:
After training the model, we need to get the final model for testing and inference. In every epoch, the model checkpoint is saved, so we can choose the best model from them based on the validation loss or we can sort them and average the parameters of the top-k models to get the final model. We can use stage 2 to do this, and the code is shown below:
```bash
if [ ${stage} -le 2 ] && [ ${stop_stage} -ge 2 ]; then
# avg n best model
avg.sh best exp/${ckpt}/checkpoints ${avg_num}
fi
```
The ```avg.sh``` is in the ```../../../utils/``` which is define in the ```path.sh```.
If you want to get the final model, you can use the script below to execute stage 0, stage 1, and stage 2:
You need to prepare an audio file, please confirm the sample rate of the audio is 16K. Assume the path of the audio file is ```data/test_audio.wav```, you can get the result by running the script below.
It is not supported at now, so we set a large stage number for this stage.
## Stage: 101 Language model training and TLG building (Need further installation! )
You need to install the kaldi and srilm to use stage 101, it is used for training language model and building TLG. To do further installation, you need to do these:
You need to prepare an audio file, please confirm the sample rate of the audio is 16K. Assume the path of the audio file is ```data/test_audio.wav```, you can get the result by running the script below.