README.md

---
license: other
tags:
- image-classification
- text-to-image
- diffusers
pretty_name: CatDataset1k
language:
- en
---
# CatDataset1k

Dataset of **1000 images of cats** (domestic cats, `Felis catus`) for training models,
experiments and fine-tuning (image generation, classification, etc.).

- Query: `cat`
- Caption / label for every image: `cat`
- Files: `cat_0000.jpg` ... `cat_0999.jpg` (JPEG)
- Sources: Wikimedia Commons + Flickr (via Openverse), open licenses

## How to download / use

### 1. Load directly with the datasets library (recommended)

```python
from datasets import load_dataset

ds = load_dataset("debugdll/DataCat1k")
# ds["train"][0]["image"]  -> PIL image
# ds["train"][0]["text"]   -> "cat"
```

No token required — the dataset is public. Total size ~ a few hundred MB.

Streaming (no full download, images loaded on demand):

```python
ds = load_dataset("debugdll/DataCat1k", streaming=True)
row = next(iter(ds["train"]))
```

### 2. Download the files with the CLI

```bash
pip install huggingface_hub
huggingface-cli download debugdll/DataCat1k
```

### 3. Download with git

```bash
git clone https://huggingface.co/datasets/debugdll/DataCat1k
```

### 4. Download individual images (browser / direct link)

```
https://huggingface.co/datasets/debugdll/DataCat1k/resolve/main/cat_0000.jpg
```

Change the filename `cat_0000.jpg` in the link to get any other image.

## Format

- `metadata.csv` — columns `file_name,text`
- The CSV + images use the standard Hugging Face `imagefolder` layout,
  so `load_dataset("debugdll/DataCat1k")` is inferred automatically.

## Training usage (diffusers LoRA / text-to-image)

```python
from datasets import load_dataset

ds = load_dataset("debugdll/DataCat1k", split="train")  # column: image, text="cat"
```

## Notes

- Caption for every image: `cat`
- All images are public-domain / openly licensed photos