Datasets:

TLVR
/

lvbench_qastyle_formmated_as_longvale_250927

dataset_info:
  features:
    - name: video_id
      dtype: string
    - name: duration
      dtype: string
    - name: timestamp
      list: float64
    - name: sentence
      dtype: string
    - name: source
      dtype: string
  splits:
    - name: train
      num_bytes: 269044
      num_examples: 1372
  download_size: 95745
  dataset_size: 269044
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*

sentence가 qa형태. 아직 description으로 바꾸기 전.

sentence_final = f"{question_text}\nAnswer:{correct_answer_text}\nQuestion Type:{question_type_text}"

DatasetDict({
    train: Dataset({
        features: ['video_id', 'duration', 'timestamp', 'sentence', 'source'],
        num_rows: 1372
    })
})

import datasets
ds_load_check = datasets.load_dataset('TLVR/lvbench_qastyle_formmated_as_longvale_250927')