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)