metadata
dataset_info:
features:
- name: urdu_instruction
dtype: string
- name: urdu_input
dtype: string
- name: urdu_output
dtype: string
- name: prompt
dtype: string
- name: input_ids
sequence: int64
- name: attention_mask
sequence: int64
splits:
- name: train
num_bytes: 60075161
num_examples: 5782
download_size: 12595600
dataset_size: 60075161
configs:
- config_name: default
data_files:
- split: train
path: data/train-*