|
|
--- |
|
|
annotations_creators: |
|
|
- manually-created |
|
|
language: |
|
|
- en |
|
|
license: apache-2.0 |
|
|
multilinguality: monolingual |
|
|
size_categories: |
|
|
- n<1K |
|
|
source_datasets: |
|
|
- original |
|
|
task_categories: |
|
|
- image-classification |
|
|
- text-to-image |
|
|
pretty_name: My Face Dataset |
|
|
dataset_info: |
|
|
features: [] |
|
|
splits: |
|
|
- name: train |
|
|
num_bytes: 0 |
|
|
num_examples: 0 |
|
|
download_size: 324 |
|
|
dataset_size: 0 |
|
|
configs: |
|
|
- config_name: default |
|
|
data_files: |
|
|
- split: train |
|
|
path: data/train-* |
|
|
--- |
|
|
|
|
|
# My Face Dataset |
|
|
|
|
|
This dataset contains a collection of face images with corresponding text descriptions (prompts). It is designed for applications such as: |
|
|
|
|
|
- Fine-tuning diffusion models (e.g., Stable Diffusion, LoRA) |
|
|
- Face recognition |
|
|
- Text-to-image generation |
|
|
|
|
|
## 🧾 Dataset Structure |
|
|
|
|
|
- `train/` — folder containing all the `.jpg` image files. |
|
|
- `metadata.jsonl` — a JSON Lines file where each line is a JSON object with the following fields: |
|
|
- `file_name`: name of the image file. |
|
|
- `text`: a description (prompt) of the image. |
|
|
|
|
|
Example entry: |
|
|
|
|
|
```json |
|
|
{"file_name": "0001.jpg", "text": "A front-facing portrait of a man smiling with natural lighting"} |
|
|
|
