Datasets:

Modalities:
Tabular
Text
Formats:
parquet
ArXiv:
Libraries:
Datasets
pandas
planbench / README.md
sileod's picture
Upload README.md with huggingface_hub
259b179 verified
|
raw
history blame
1.31 kB
metadata
dataset_info:
  - config_name: task_1_plan_generation
    features:
      - name: task
        dtype: string
      - name: prompt_type
        dtype: string
      - name: domain
        dtype: string
      - name: instance_id
        dtype: int64
      - name: example_instance_ids
        sequence: int64
      - name: query
        dtype: string
      - name: ground_truth_plan
        dtype: string
    splits:
      - name: train
        num_bytes: 8849445
        num_examples: 2270
    download_size: 1029082
    dataset_size: 8849445
  - config_name: task_2_plan_optimality
    features:
      - name: task
        dtype: string
      - name: prompt_type
        dtype: string
      - name: domain
        dtype: string
      - name: instance_id
        dtype: int64
      - name: example_instance_ids
        sequence: int64
      - name: query
        dtype: string
      - name: ground_truth_plan
        dtype: string
    splits:
      - name: train
        num_bytes: 6387505
        num_examples: 1692
    download_size: 0
    dataset_size: 6387505
configs:
  - config_name: task_1_plan_generation
    data_files:
      - split: train
        path: task_1_plan_generation/train-*
  - config_name: task_2_plan_optimality
    data_files:
      - split: train
        path: task_2_plan_optimality/train-*

Dataset Card for "planbench"

More Information needed