testds

license: Apache License 2.0
tasks:
- text-generation
image:
image-classification:
size_scale:
- 0-100

## 数据集描述
Uni-Fold-Data 开源的蛋白质折叠训练数据。

### 数据集简介
该数据集用于蛋白质折叠模型训练，参考： [Uni-Fold](https://github.com/dptech-corp/Uni-Fold).

### 数据集支持的任务
该数据集适用于Uni-Fold-Multimer蛋白质复合物结构预测模型，以及Uni-Fold-Monomer蛋白质单体结构预测模型，以上模型均已在modelscope社区开放。

## 数据集的格式和结构

### 数据格式
包含多个文件夹，每个文件夹下包含多个gz压缩文件，解压后为pickle格式。该数据集全量下载要求至少3TB的存储空间，用于保存压缩文件和解压后的pickle文件。

### Clone with HTTP
```bash
git clone https://www.modelscope.cn/datasets/joannyli/testds.git
```

license: Apache License 2.0
tasks:
- text-generation
image:
image-classification:
size_scale:
- 0-100

## Dataset Description
Uni-Fold-Data is an open-sourced training dataset for protein folding.

### Dataset Overview
This dataset is designed for training protein folding models, with reference to [Uni-Fold](https://github.com/dptech-corp/Uni-Fold).

### Supported Tasks
This dataset is compatible with both the Uni-Fold-Multimer protein complex structure prediction model and the Uni-Fold-Monomer protein monomer structure prediction model. Both of these models have been publicly released on the ModelScope community.

## Dataset Format and Structure
### Data Format
The dataset consists of multiple directories, each containing several gzip-compressed (.gz) files. After decompression, the files are in pickle format. A full download of this dataset requires at least 3TB of storage space to store both the compressed files and the decompressed pickle files.

### Clone with HTTP
bash
git clone https://www.modelscope.cn/datasets/joannyli/testds.git