recbole.data.dataloader.neg_sample_mixin

class recbole.data.dataloader.neg_sample_mixin.NegSampleByMixin(config, dataset, sampler, neg_sample_args, batch_size=1, dl_format=<InputType.POINTWISE: 1>, shuffle=False)[source]

Bases: recbole.data.dataloader.neg_sample_mixin.NegSampleMixin

NegSampleByMixin is an abstract class which can sample negative examples by ratio. It has two neg-sampling method, the one is 1-by-1 neg-sampling (pair wise), and the other is 1-by-multi neg-sampling (point wise).

Parameters

  • config (Config) – The config of dataloader.

  • dataset (Dataset) – The dataset of dataloader.

  • sampler (Sampler) – The sampler of dataloader.

  • neg_sample_args (dict) – The neg_sample_args of dataloader.

  • batch_size (int, optional) – The batch_size of dataloader. Defaults to 1.

  • dl_format (InputType, optional) – The input type of dataloader. Defaults to POINTWISE.

  • shuffle (bool, optional) – Whether the dataloader will be shuffle after a round. Defaults to False.

class recbole.data.dataloader.neg_sample_mixin.NegSampleMixin(config, dataset, sampler, neg_sample_args, batch_size=1, dl_format=<InputType.POINTWISE: 1>, shuffle=False)[source]

Bases: recbole.data.dataloader.abstract_dataloader.AbstractDataLoader

NegSampleMixin is a abstract class, all dataloaders that need negative sampling should inherit this class. This class provides some necessary parameters and method for negative sampling, such as neg_sample_args and _neg_sampling() and so on.

Parameters

  • config (Config) – The config of dataloader.

  • dataset (Dataset) – The dataset of dataloader.

  • sampler (Sampler) – The sampler of dataloader.

  • neg_sample_args (dict) – The neg_sample_args of dataloader.

  • batch_size (int, optional) – The batch_size of dataloader. Defaults to 1.

  • dl_format (InputType, optional) – The input type of dataloader. Defaults to POINTWISE.

  • shuffle (bool, optional) – Whether the dataloader will be shuffle after a round. Defaults to False.

data_preprocess()[source]

Do neg-sampling before training/evaluation.

dl_type = 3
get_pos_len_list()[source]
Returns

Number of positive item for each user in a training/evaluating epoch.

Return type

np.ndarray or list

setup()[source]

Do batch size adaptation.