| --- |
| license: apache-2.0 |
| dataset_info: |
| features: |
| - name: smiles |
| dtype: string |
| - name: activity |
| dtype: float64 |
| splits: |
| - name: train |
| num_bytes: 13736 |
| num_examples: 513 |
| - name: validation |
| num_bytes: 1721 |
| num_examples: 64 |
| - name: test |
| num_bytes: 1780 |
| num_examples: 65 |
| download_size: 15823 |
| dataset_size: 17237 |
| configs: |
| - config_name: default |
| data_files: |
| - split: train |
| path: data/train-* |
| - split: validation |
| path: data/validation-* |
| - split: test |
| path: data/test-* |
| --- |
| The dataset is created using data from the deepchem library introduced by moleculenet. |
| The dataset is split in a random fashion and normalized using z_score. |
| The random seed used for splitting is 123. |
| The smiles are canonicalized using RDkit. |
