recbole.model.init

recbole.model.init.xavier_normal_initialization(module)[source]

using xavier_normal_ in PyTorch to initialize the parameters in nn.Embedding and nn.Linear layers. For bias in nn.Linear layers, using constant 0 to initialize.

Examples

>>> self.apply(xavier_normal_initialization)
recbole.model.init.xavier_uniform_initialization(module)[source]

using xavier_uniform_ in PyTorch to initialize the parameters in nn.Embedding and nn.Linear layers. For bias in nn.Linear layers, using constant 0 to initialize.

Examples

>>> self.apply(xavier_uniform_initialization)