Datasets:

Modalities:
Text
Formats:
parquet
Libraries:
Datasets
pandas
rxn-dataset / README.md
LioNoussi's picture
Separated input_ids and labels again. Removed attention mask.
d0900cd verified
|
raw
history blame
538 Bytes
metadata
dataset_info:
  features:
    - name: input_ids
      dtype: string
    - name: labels
      dtype: string
  splits:
    - name: train
      num_bytes: 95684951
      num_examples: 409035
    - name: validation
      num_bytes: 7001813
      num_examples: 30000
    - name: test
      num_bytes: 9377152
      num_examples: 40000
  download_size: 44425015
  dataset_size: 112063916
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: validation
        path: data/validation-*
      - split: test
        path: data/test-*