| # Summarization Fine-tuning Dataset | |
| A dataset of 2000 examples for fine-tuning small language models on summarization tasks. | |
| ## Statistics | |
| - **Total examples**: 2000 | |
| - **Train examples**: 1600 (80.0%) | |
| - **Validation examples**: 200 (10.0%) | |
| - **Test examples**: 200 (10.0%) | |
| ## Dataset Distribution | |
| | Dataset | Count | Percentage | | |
| |---------|-------|------------| | |
| | xsum | 2000 | 100.0% | | |
| ## Format | |
| The dataset is provided in alpaca format. | |
| ## Configuration | |
| - **Maximum tokens**: 2000 | |
| - **Tokenizer**: gpt2 | |
| - **Random seed**: 42 | |
| ## Usage | |
| ```python | |
| from datasets import load_dataset | |
| # Load the dataset | |
| dataset = load_dataset("YOUR_USERNAME/summarization-finetune-10k") | |
| # Access the splits | |
| train_data = dataset["train"] | |
| val_data = dataset["validation"] | |
| test_data = dataset["test"] | |
| ``` | |