README.md

metadata

language: []
pretty_name: alagesse/gnn-datasets
configs:
  - config_name: zinc-12k
    data_files:
      - split: train
        path: zinc-12k/train*.parquet
      - split: validation
        path: zinc-12k/validation*.parquet
      - split: test
        path: zinc-12k/test*.parquet
  - config_name: zinc-250k
    data_files:
      - split: train
        path: zinc-250k/train*.parquet
      - split: validation
        path: zinc-250k/validation*.parquet
      - split: test
        path: zinc-250k/test*.parquet
  - config_name: aqsol
    data_files:
      - split: train
        path: aqsol/train*.parquet
      - split: validation
        path: aqsol/validation*.parquet
      - split: test
        path: aqsol/test*.parquet
  - config_name: pcqm4mv2-2d
    data_files:
      - split: train
        path: pcqm4mv2-2d/train*.parquet
      - split: validation
        path: pcqm4mv2-2d/validation*.parquet
      - split: test_dev
        path: pcqm4mv2-2d/test_dev*.parquet
      - split: test_challenge
        path: pcqm4mv2-2d/test_challenge*.parquet
  - config_name: pcqm4mv2-3d
    data_files:
      - split: train
        path: pcqm4mv2-3d/train*.parquet
  - config_name: pcqm4mv2-smiles
    data_files:
      - split: train
        path: pcqm4mv2-smiles/train*.parquet
      - split: validation
        path: pcqm4mv2-smiles/validation*.parquet
      - split: test_dev
        path: pcqm4mv2-smiles/test_dev*.parquet
      - split: test_challenge
        path: pcqm4mv2-smiles/test_challenge*.parquet
  - config_name: qm9
    data_files:
      - split: train
        path: qm9/train*.parquet
      - split: validation
        path: qm9/validation*.parquet
      - split: test
        path: qm9/test*.parquet
  - config_name: qm9-smiles
    data_files:
      - split: train
        path: qm9-smiles/train*.parquet
      - split: validation
        path: qm9-smiles/validation*.parquet
      - split: test
        path: qm9-smiles/test*.parquet

alagesse/gnn-datasets

This repository hosts multiple standardized graph datasets. Load any configuration via:

from datasets import load_dataset
ds = load_dataset("alagesse/gnn-datasets", "<config_name>", split="<split_name>")

Available configurations:

Configuration
zinc-12k
zinc-250k
aqsol
pcqm4mv2-2d
pcqm4mv2-3d
pcqm4mv2-smiles
qm9
qm9-smiles