# Jasper

**Repository Path**: echo_lin/Jasper

## Basic Information

- **Project Name**: Jasper
- **Description**: No description available
- **Primary Language**: Python
- **License**: Apache-2.0
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2021-02-23
- **Last Updated**: 2021-03-02

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README


# Contents
- [Contents](#contents)
- [Description](#bert-description)
- [Model Architecture](#model-architecture)
- [Dataset](#dataset)
- [Environment Requirements](#environment-requirements)
- [Quick Start](#quick-start)

# [Description](#contents)
Jasper is a family of end-to-end ASR models that replace acoustic and pronunciation models with a convolutional neural network.


[Paper](https://arxiv.org/abs/1904.03288): Jason Li, Vitaly Lavrukhin, Boris Ginsburg, Ryan Leary, Oleksii Kuchaiev, Jonathan M. Cohen, Huyen Nguyen, Ravi Teja Gaddecc. [Jasper: An End-to-End Convolutional Neural Acoustic Model.]((https://arxiv.org/abs/1904.03288)).

# [Model Architecture](#contents)
Jasper uses mel-filterbank features calculated from 20ms
windows with a 10ms overlap, and outputs a probability distribution over characters per frame2
. Jasper has a block architecture: a Jasper BxR model has B blocks, each with R subblocks. Each sub-block applies the following operations: a 1Dconvolution, batch norm, ReLU, and dropout. All sub-blocks in
a block have the same number of output channels.

# [Dataset](#contents)
- LibriSpeech is a ~1,000 hours of 16kHz read English speech corpus. The data is obtained from audiobooks read from the LibriVox project, and has been segmented and aligned.

# [Environment Requirements](#contents)
- Hardware（Ascend）
  - Prepare hardware environment with Ascend910 processor. If you want to try Ascend, please send the [application form](https://obs-9be7.obs.cn-east-2.myhuaweicloud.com/file/other/Ascend%20Model%20Zoo%E4%BD%93%E9%AA%8C%E8%B5%84%E6%BA%90%E7%94%B3%E8%AF%B7%E8%A1%A8.docx) to ascend@huawei.com. Once approved, you can get access to the resources.
- Framework
  - [Tensorflow 1.15.0](https://www.tensorflow.org/versions/)
- Requirement module if necessary
```bash
# run pip to install reqirement
pip install -r requirement.txt
```
# [Quick Start](#contents)
How can we start training and evaluation as soon as possible. One tutorial has been provided on ModelArts Platform with Pycharm IDE and its ToolKit. You can refer to the [tutorial link](https://gitee.com/echo_lin/Jasper/blob/master/README_TUTORIAL.md) for more information.