pokemon-blip-captions

# Dataset Card for Pokémon BLIP captions

_Dataset used to train [Pokémon text to image model](https://github.com/LambdaLabsML/examples/tree/main/stable-diffusion-finetuning)_

BLIP generated captions for Pokémon images from Few Shot Pokémon dataset introduced by _Towards Faster and Stabilized GAN Training for High-fidelity Few-shot Image Synthesis_ (FastGAN). Original images were obtained from [FastGAN-pytorch](https://github.com/odegeasslbc/FastGAN-pytorch) and captioned with the [pre-trained BLIP model](https://github.com/salesforce/BLIP).

For each row the dataset contains `image` and `text` keys. `image` is a varying size PIL jpeg, and `text` is the accompanying text caption. Only a train split is provided.

## 下载数据集
```
git clone http://oauth2:[e-mail]/datasets/AI-ModelScope/pokemon-blip-captions.git
```

## 示例代码
```python
from modelscope import MsDataset
from modelscope.utils.constant import DownloadMode

ds = MsDataset.load('AI-ModelScope/pokemon-blip-captions',subset_name='default', split='train', download_mode=DownloadMode.FORCE_REDOWNLOAD, cache_dir='/mnt/workspace/cache_yk')
print(next(iter(ds)))
```

## Examples

> a drawing of a green pokemon with red eyes

> a green and yellow toy with a red nose

> a red and white ball with an angry look on its face

# 宝可梦BLIP标注数据集卡片
_本数据集用于训练[宝可梦文本到图像模型](https://github.com/LambdaLabsML/examples/tree/main/stable-diffusion-finetuning)_

BLIP生成的标注文本来自由《Towards Faster and Stabilized GAN Training for High-fidelity Few-shot Image Synthesis》（FastGAN）提出的少样本宝可梦数据集。原始图像取自[FastGAN-pytorch](https://github.com/odegeasslbc/FastGAN-pytorch)，并通过[预训练BLIP模型](https://github.com/salesforce/BLIP)生成标注。

数据集中每一条目均包含`image`与`text`两个键：`image`为尺寸可变的PIL格式JPEG图像，`text`为对应的文本标注。本数据集仅提供训练拆分。

## 数据集下载

git clone http://oauth2:[e-mail]/datasets/AI-ModelScope/pokemon-blip-captions.git

## 示例代码
python
from modelscope import MsDataset
from modelscope.utils.constant import DownloadMode

ds = MsDataset.load('AI-ModelScope/pokemon-blip-captions',subset_name='default', split='train', download_mode=DownloadMode.FORCE_REDOWNLOAD, cache_dir='/mnt/workspace/cache_yk')
print(next(iter(ds)))

## 示例
> 一幅带有红色眼睛的绿色宝可梦绘图

> 一个带有红色鼻子的绿黄色玩具

> 一个表面带有愤怒表情的红白配色球体

搜集汇总

背景与挑战

背景概述

该数据集用于训练宝可梦文本到图像模型，基于Few Shot Pokémon数据集，通过BLIP模型自动生成图像描述。数据集包含图像和对应的文本描述键值对，仅提供训练分割，适用于图像生成任务的训练。