| --- |
| license: ecl-2.0 |
| --- |
| # Repo |
| git@hf.co:datasets/telcom/deewaiREALCN-training |
|
|
| # DeewaiREALCN Training Data |
|
|
| Image–text pairs for training captioning or vision–language models. Each image is a 1024×1024 RGB JPEG portrait with a short English description. |
|
|
| ## Contents |
| - `data/train/`: 9,000 pairs for training. |
| - `images/`: JPEG files (`090000.jpg`, …). |
| - `captions.jsonl`: one JSON object per line with `file_name` and `text`. |
| - `data/val/`: 1,000 pairs for validation with the same layout. |
|
|
| Example entry: |
| ```json |
| {"file_name": "090000.jpg", "text": "smiling woman with long brown hair and a white shirt"} |
| ``` |
|
|
| ## Usage |
| 1. Read `captions.jsonl` line by line (JSON Lines format). |
| 2. Join on `file_name` to locate the matching image in the corresponding `images/` folder. |
| 3. Use `data/train` for model fitting and `data/val` for evaluation. |
|
|
| ## Notes |
| - Images are 1024×1024 RGB. |
| - License: Eclipse Public License 2.0 (see header above). |
|
|
