Google I/O — это обертка! Наверстать упущенное в сеансах TensorFlow Просмотреть сеансы

mslr_web

  • Описание :

MSLR-WEB — это два крупномасштабных набора данных Learning-to-Rank, выпущенных Microsoft Research. Первый набор данных (называемый «30 000») содержит 30 000 запросов, а второй набор данных (называемый «10 000») содержит 10 000 запросов. Каждый набор данных состоит из пар запрос-документ, представленных в виде векторов признаков и соответствующих меток суждения о релевантности.

Вы можете указать, использовать ли версию набора данных «10 КБ» или «30 КБ» и соответствующую складку следующим образом:

ds = tfds.load("mslr_web/30k_fold1")

Если указан только mslr_web , по умолчанию выбирается вариант mslr_web/10k_fold1 :

# This is the same as `tfds.load("mslr_web/10k_fold1")`
ds = tfds.load("mslr_web")
FeaturesDict({
    'bm25_anchor': Tensor(shape=(None,), dtype=tf.float64),
    'bm25_body': Tensor(shape=(None,), dtype=tf.float64),
    'bm25_title': Tensor(shape=(None,), dtype=tf.float64),
    'bm25_url': Tensor(shape=(None,), dtype=tf.float64),
    'bm25_whole_document': Tensor(shape=(None,), dtype=tf.float64),
    'boolean_model_anchor': Tensor(shape=(None,), dtype=tf.float64),
    'boolean_model_body': Tensor(shape=(None,), dtype=tf.float64),
    'boolean_model_title': Tensor(shape=(None,), dtype=tf.float64),
    'boolean_model_url': Tensor(shape=(None,), dtype=tf.float64),
    'boolean_model_whole_document': Tensor(shape=(None,), dtype=tf.float64),
    'covered_query_term_number_anchor': Tensor(shape=(None,), dtype=tf.float64),
    'covered_query_term_number_body': Tensor(shape=(None,), dtype=tf.float64),
    'covered_query_term_number_title': Tensor(shape=(None,), dtype=tf.float64),
    'covered_query_term_number_url': Tensor(shape=(None,), dtype=tf.float64),
    'covered_query_term_number_whole_document': Tensor(shape=(None,), dtype=tf.float64),
    'covered_query_term_ratio_anchor': Tensor(shape=(None,), dtype=tf.float64),
    'covered_query_term_ratio_body': Tensor(shape=(None,), dtype=tf.float64),
    'covered_query_term_ratio_title': Tensor(shape=(None,), dtype=tf.float64),
    'covered_query_term_ratio_url': Tensor(shape=(None,), dtype=tf.float64),
    'covered_query_term_ratio_whole_document': Tensor(shape=(None,), dtype=tf.float64),
    'idf_anchor': Tensor(shape=(None,), dtype=tf.float64),
    'idf_body': Tensor(shape=(None,), dtype=tf.float64),
    'idf_title': Tensor(shape=(None,), dtype=tf.float64),
    'idf_url': Tensor(shape=(None,), dtype=tf.float64),
    'idf_whole_document': Tensor(shape=(None,), dtype=tf.float64),
    'inlink_number': Tensor(shape=(None,), dtype=tf.float64),
    'label': Tensor(shape=(None,), dtype=tf.float64),
    'length_of_url': Tensor(shape=(None,), dtype=tf.float64),
    'lmir_abs_anchor': Tensor(shape=(None,), dtype=tf.float64),
    'lmir_abs_body': Tensor(shape=(None,), dtype=tf.float64),
    'lmir_abs_title': Tensor(shape=(None,), dtype=tf.float64),
    'lmir_abs_url': Tensor(shape=(None,), dtype=tf.float64),
    'lmir_abs_whole_document': Tensor(shape=(None,), dtype=tf.float64),
    'lmir_dir_anchor': Tensor(shape=(None,), dtype=tf.float64),
    'lmir_dir_body': Tensor(shape=(None,), dtype=tf.float64),
    'lmir_dir_title': Tensor(shape=(None,), dtype=tf.float64),
    'lmir_dir_url': Tensor(shape=(None,), dtype=tf.float64),
    'lmir_dir_whole_document': Tensor(shape=(None,), dtype=tf.float64),
    'lmir_jm_anchor': Tensor(shape=(None,), dtype=tf.float64),
    'lmir_jm_body': Tensor(shape=(None,), dtype=tf.float64),
    'lmir_jm_title': Tensor(shape=(None,), dtype=tf.float64),
    'lmir_jm_url': Tensor(shape=(None,), dtype=tf.float64),
    'lmir_jm_whole_document': Tensor(shape=(None,), dtype=tf.float64),
    'max_of_stream_length_normalized_term_frequency_anchor': Tensor(shape=(None,), dtype=tf.float64),
    'max_of_stream_length_normalized_term_frequency_body': Tensor(shape=(None,), dtype=tf.float64),
    'max_of_stream_length_normalized_term_frequency_title': Tensor(shape=(None,), dtype=tf.float64),
    'max_of_stream_length_normalized_term_frequency_url': Tensor(shape=(None,), dtype=tf.float64),
    'max_of_stream_length_normalized_term_frequency_whole_document': Tensor(shape=(None,), dtype=tf.float64),
    'max_of_term_frequency_anchor': Tensor(shape=(None,), dtype=tf.float64),
    'max_of_term_frequency_body': Tensor(shape=(None,), dtype=tf.float64),
    'max_of_term_frequency_title': Tensor(shape=(None,), dtype=tf.float64),
    'max_of_term_frequency_url': Tensor(shape=(None,), dtype=tf.float64),
    'max_of_term_frequency_whole_document': Tensor(shape=(None,), dtype=tf.float64),
    'max_of_tf_idf_anchor': Tensor(shape=(None,), dtype=tf.float64),
    'max_of_tf_idf_body': Tensor(shape=(None,), dtype=tf.float64),
    'max_of_tf_idf_title': Tensor(shape=(None,), dtype=tf.float64),
    'max_of_tf_idf_url': Tensor(shape=(None,), dtype=tf.float64),
    'max_of_tf_idf_whole_document': Tensor(shape=(None,), dtype=tf.float64),
    'mean_of_stream_length_normalized_term_frequency_anchor': Tensor(shape=(None,), dtype=tf.float64),
    'mean_of_stream_length_normalized_term_frequency_body': Tensor(shape=(None,), dtype=tf.float64),
    'mean_of_stream_length_normalized_term_frequency_title': Tensor(shape=(None,), dtype=tf.float64),
    'mean_of_stream_length_normalized_term_frequency_url': Tensor(shape=(None,), dtype=tf.float64),
    'mean_of_stream_length_normalized_term_frequency_whole_document': Tensor(shape=(None,), dtype=tf.float64),
    'mean_of_term_frequency_anchor': Tensor(shape=(None,), dtype=tf.float64),
    'mean_of_term_frequency_body': Tensor(shape=(None,), dtype=tf.float64),
    'mean_of_term_frequency_title': Tensor(shape=(None,), dtype=tf.float64),
    'mean_of_term_frequency_url': Tensor(shape=(None,), dtype=tf.float64),
    'mean_of_term_frequency_whole_document': Tensor(shape=(None,), dtype=tf.float64),
    'mean_of_tf_idf_anchor': Tensor(shape=(None,), dtype=tf.float64),
    'mean_of_tf_idf_body': Tensor(shape=(None,), dtype=tf.float64),
    'mean_of_tf_idf_title': Tensor(shape=(None,), dtype=tf.float64),
    'mean_of_tf_idf_url': Tensor(shape=(None,), dtype=tf.float64),
    'mean_of_tf_idf_whole_document': Tensor(shape=(None,), dtype=tf.float64),
    'min_of_stream_length_normalized_term_frequency_anchor': Tensor(shape=(None,), dtype=tf.float64),
    'min_of_stream_length_normalized_term_frequency_body': Tensor(shape=(None,), dtype=tf.float64),
    'min_of_stream_length_normalized_term_frequency_title': Tensor(shape=(None,), dtype=tf.float64),
    'min_of_stream_length_normalized_term_frequency_url': Tensor(shape=(None,), dtype=tf.float64),
    'min_of_stream_length_normalized_term_frequency_whole_document': Tensor(shape=(None,), dtype=tf.float64),
    'min_of_term_frequency_anchor': Tensor(shape=(None,), dtype=tf.float64),
    'min_of_term_frequency_body': Tensor(shape=(None,), dtype=tf.float64),
    'min_of_term_frequency_title': Tensor(shape=(None,), dtype=tf.float64),
    'min_of_term_frequency_url': Tensor(shape=(None,), dtype=tf.float64),
    'min_of_term_frequency_whole_document': Tensor(shape=(None,), dtype=tf.float64),
    'min_of_tf_idf_anchor': Tensor(shape=(None,), dtype=tf.float64),
    'min_of_tf_idf_body': Tensor(shape=(None,), dtype=tf.float64),
    'min_of_tf_idf_title': Tensor(shape=(None,), dtype=tf.float64),
    'min_of_tf_idf_url': Tensor(shape=(None,), dtype=tf.float64),
    'min_of_tf_idf_whole_document': Tensor(shape=(None,), dtype=tf.float64),
    'number_of_slash_in_url': Tensor(shape=(None,), dtype=tf.float64),
    'outlink_number': Tensor(shape=(None,), dtype=tf.float64),
    'page_rank': Tensor(shape=(None,), dtype=tf.float64),
    'quality_score': Tensor(shape=(None,), dtype=tf.float64),
    'quality_score_2': Tensor(shape=(None,), dtype=tf.float64),
    'query_url_click_count': Tensor(shape=(None,), dtype=tf.float64),
    'site_rank': Tensor(shape=(None,), dtype=tf.float64),
    'stream_length_anchor': Tensor(shape=(None,), dtype=tf.float64),
    'stream_length_body': Tensor(shape=(None,), dtype=tf.float64),
    'stream_length_title': Tensor(shape=(None,), dtype=tf.float64),
    'stream_length_url': Tensor(shape=(None,), dtype=tf.float64),
    'stream_length_whole_document': Tensor(shape=(None,), dtype=tf.float64),
    'sum_of_stream_length_normalized_term_frequency_anchor': Tensor(shape=(None,), dtype=tf.float64),
    'sum_of_stream_length_normalized_term_frequency_body': Tensor(shape=(None,), dtype=tf.float64),
    'sum_of_stream_length_normalized_term_frequency_title': Tensor(shape=(None,), dtype=tf.float64),
    'sum_of_stream_length_normalized_term_frequency_url': Tensor(shape=(None,), dtype=tf.float64),
    'sum_of_stream_length_normalized_term_frequency_whole_document': Tensor(shape=(None,), dtype=tf.float64),
    'sum_of_term_frequency_anchor': Tensor(shape=(None,), dtype=tf.float64),
    'sum_of_term_frequency_body': Tensor(shape=(None,), dtype=tf.float64),
    'sum_of_term_frequency_title': Tensor(shape=(None,), dtype=tf.float64),
    'sum_of_term_frequency_url': Tensor(shape=(None,), dtype=tf.float64),
    'sum_of_term_frequency_whole_document': Tensor(shape=(None,), dtype=tf.float64),
    'sum_of_tf_idf_anchor': Tensor(shape=(None,), dtype=tf.float64),
    'sum_of_tf_idf_body': Tensor(shape=(None,), dtype=tf.float64),
    'sum_of_tf_idf_title': Tensor(shape=(None,), dtype=tf.float64),
    'sum_of_tf_idf_url': Tensor(shape=(None,), dtype=tf.float64),
    'sum_of_tf_idf_whole_document': Tensor(shape=(None,), dtype=tf.float64),
    'url_click_count': Tensor(shape=(None,), dtype=tf.float64),
    'url_dwell_time': Tensor(shape=(None,), dtype=tf.float64),
    'variance_of_stream_length_normalized_term_frequency_anchor': Tensor(shape=(None,), dtype=tf.float64),
    'variance_of_stream_length_normalized_term_frequency_body': Tensor(shape=(None,), dtype=tf.float64),
    'variance_of_stream_length_normalized_term_frequency_title': Tensor(shape=(None,), dtype=tf.float64),
    'variance_of_stream_length_normalized_term_frequency_url': Tensor(shape=(None,), dtype=tf.float64),
    'variance_of_stream_length_normalized_term_frequency_whole_document': Tensor(shape=(None,), dtype=tf.float64),
    'variance_of_term_frequency_anchor': Tensor(shape=(None,), dtype=tf.float64),
    'variance_of_term_frequency_body': Tensor(shape=(None,), dtype=tf.float64),
    'variance_of_term_frequency_title': Tensor(shape=(None,), dtype=tf.float64),
    'variance_of_term_frequency_url': Tensor(shape=(None,), dtype=tf.float64),
    'variance_of_term_frequency_whole_document': Tensor(shape=(None,), dtype=tf.float64),
    'variance_of_tf_idf_anchor': Tensor(shape=(None,), dtype=tf.float64),
    'variance_of_tf_idf_body': Tensor(shape=(None,), dtype=tf.float64),
    'variance_of_tf_idf_title': Tensor(shape=(None,), dtype=tf.float64),
    'variance_of_tf_idf_url': Tensor(shape=(None,), dtype=tf.float64),
    'variance_of_tf_idf_whole_document': Tensor(shape=(None,), dtype=tf.float64),
    'vector_space_model_anchor': Tensor(shape=(None,), dtype=tf.float64),
    'vector_space_model_body': Tensor(shape=(None,), dtype=tf.float64),
    'vector_space_model_title': Tensor(shape=(None,), dtype=tf.float64),
    'vector_space_model_url': Tensor(shape=(None,), dtype=tf.float64),
    'vector_space_model_whole_document': Tensor(shape=(None,), dtype=tf.float64),
})
@article{DBLP:journals/corr/QinL13,
  author    = {Tao Qin and Tie{-}Yan Liu},
  title     = {Introducing {LETOR} 4.0 Datasets},
  journal   = {CoRR},
  volume    = {abs/1306.2597},
  year      = {2013},
  url       = {http://arxiv.org/abs/1306.2597},
  timestamp = {Mon, 01 Jul 2013 20:31:25 +0200},
  biburl    = {http://dblp.uni-trier.de/rec/bib/journals/corr/QinL13},
  bibsource = {dblp computer science bibliography, http://dblp.org}
}

mslr_web/10k_fold1 (конфигурация по умолчанию)

  • Размер загрузки : 1.15 GiB

  • Размер набора данных : 381.58 MiB .

  • Сплиты :

Расколоть Примеры
'test' 2000
'train' 6000
'vali' 2000

mslr_web/10k_fold2

  • Размер загрузки : 1.15 GiB

  • Размер набора данных : 381.58 MiB .

  • Сплиты :

Расколоть Примеры
'test' 2000
'train' 6000
'vali' 2000

mslr_web/10k_fold3

  • Размер загрузки : 1.15 GiB

  • Размер набора данных : 381.58 MiB .

  • Сплиты :

Расколоть Примеры
'test' 2000
'train' 6000
'vali' 2000

mslr_web/10k_fold4

  • Размер загрузки : 1.15 GiB

  • Размер набора данных : 381.58 MiB .

  • Сплиты :

Расколоть Примеры
'test' 2000
'train' 6000
'vali' 2000

mslr_web/10k_fold5

  • Размер загрузки : 1.15 GiB

  • Размер набора данных : 381.58 MiB .

  • Сплиты :

Расколоть Примеры
'test' 2000
'train' 6000
'vali' 2000

mslr_web/30k_fold1

  • Размер загрузки : 3.59 GiB

  • Размер набора данных : 1.17 GiB

  • Сплиты :

Расколоть Примеры
'test' 6306
'train' 18 919
'vali' 6306

mslr_web/30k_fold2

  • Размер загрузки : 3.59 GiB

  • Размер набора данных : 1.17 GiB

  • Сплиты :

Расколоть Примеры
'test' 6307
'train' 18 918
'vali' 6306

mslr_web/30k_fold3

  • Размер загрузки : 3.59 GiB

  • Размер набора данных : 1.17 GiB

  • Сплиты :

Расколоть Примеры
'test' 6306
'train' 18 918
'vali' 6307

mslr_web/30k_fold4

  • Размер загрузки : 3.59 GiB

  • Размер набора данных : 1.17 GiB

  • Сплиты :

Расколоть Примеры
'test' 6306
'train' 18 919
'vali' 6306

mslr_web/30k_fold5

  • Размер загрузки : 3.59 GiB

  • Размер набора данных : 1.17 GiB

  • Сплиты :

Расколоть Примеры
'test' 6306
'train' 18 919
'vali' 6306