https://github.com/deepvac/transformer

DeepVAC-compliant transformer implementation.
https://github.com/deepvac/transformer

Last synced: about 1 year ago
JSON representation

DeepVAC-compliant transformer implementation.

• Host: GitHub
• URL: https://github.com/deepvac/transformer
• Owner: DeepVAC
• License: gpl-3.0
• Created: 2021-07-27T09:53:17.000Z (almost 5 years ago)
• Default Branch: master
• Last Pushed: 2021-08-13T06:22:42.000Z (almost 5 years ago)
• Last Synced: 2025-05-09T01:07:10.701Z (about 1 year ago)
• Language: Python
• Size: 867 KB
• Stars: 5
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          
# transformer

**符合DeepVAC规范的transformer实现。**  

DeepVAC-compliant transformer implementation.

# 运行

### MLab HomePod 2.0

在MLab HomePod 2.0上运行该项目，无需安装任何依赖，只需要输入如下命令即可开始训练:

```bash

python train.py

```

### 自定义环境

需要至少安装：

- pytorch 1.8+;

- numpy

- deepvac 0.6.0+

然后输入如下命令即可开始训练：

```bash

python train.py

```

# 数据集

该项目所使用数据集来自：[ancient_morden_chinese](https://github.com/CivilNet/ancient_morden_chinese)

# 配置

该项目符合DeepVAC规范，相关配置均包含在config.py中，配置项的含义也符合DeepVAC规范，请自行阅读[DeepVAC规范](https://github.com/deepvac/deepvac)。