README.md

metadata

configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
dataset_info:
  features:
    - name: task
      dtype: string
    - name: web_name
      dtype: string
  splits:
    - name: train
      num_bytes: 16798
      num_examples: 130
  download_size: 8477
  dataset_size: 16798

Each instruction is generated from Instruction_generation_two_shot in question_generation file. These are the number of tasks for each site.

gitlab (15), openstreetmap (35), reddit (10), Magento (35), One Stop Market (35)

---

dataset_info: features: - name: task dtype: string - name: web_name dtype: string splits: - name: train num_bytes: 16798 num_examples: 130 download_size: 8477 dataset_size: 16798 configs: - config_name: default data_files: - split: train path: data/train-*