src.models.components package

Subpackages

• src.models.components.attention package
  • Submodules
  • src.models.components.attention.additive module
    • AdditiveAttention
      • AdditiveAttention.forward()
      • AdditiveAttention.training
  • src.models.components.attention.multihead_self module
    • MultiHeadSelfAttention
      • MultiHeadSelfAttention.forward()
      • MultiHeadSelfAttention.training
    • ScaledDotProductAttention
      • ScaledDotProductAttention.forward()
      • ScaledDotProductAttention.training
  • Module contents
• src.models.components.click_predictor package
  • Submodules
  • src.models.components.click_predictor.DNN module
    • DNNClickPredictor
      • DNNClickPredictor.forward()
      • DNNClickPredictor.training
  • src.models.components.click_predictor.dot_product module
    • DotProductClickPredictor
      • DotProductClickPredictor.forward()
      • DotProductClickPredictor.training
  • Module contents

Submodules

src.models.components.metrics module

class src.models.components.metrics.LogLoss(dist_sync_on_step=False)

Bases: Metric

compute()

Override this method to compute the final metric value from state variables synchronized across the distributed backend.

full_state_update: bool = False

update(preds: Tensor, target: Tensor)

Override this method to update the state variables of your metric class.

class src.models.components.metrics.MRR(dist_sync_on_step=False, k=None)

Bases: Metric

compute()

Override this method to compute the final metric value from state variables synchronized across the distributed backend.

full_state_update: bool = False

update(preds: Tensor, target: Tensor)

Override this method to update the state variables of your metric class.

class src.models.components.metrics.Senti(dist_sync_on_step=False, k=None)

Bases: Metric

compute()

Override this method to compute the final metric value from state variables synchronized across the distributed backend.

full_state_update: bool = False

update(y_pred: Tensor, s_c: Tensor, s_mean: Tensor)

Override this method to update the state variables of your metric class.

class src.models.components.metrics.SentiMRR(dist_sync_on_step=False, k=None)

Bases: Metric

compute()

Override this method to compute the final metric value from state variables synchronized across the distributed backend.

full_state_update: bool = False

update(y_pred: Tensor, s_c: Tensor, s_mean: Tensor)

Override this method to update the state variables of your metric class.

src.models.components.simple_dense_net module

class src.models.components.simple_dense_net.SimpleDenseNet(input_size: int = 784, lin1_size: int = 256, lin2_size: int = 256, lin3_size: int = 256, output_size: int = 10)

Bases: Module

forward(x)

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

training: bool

src.models.components.utils module

class src.models.components.utils.AdditiveAttention(query_dim, embedding_dim)

Bases: Module

forward(input_sequence)

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

training: bool

class src.models.components.utils.TimeDistributed(module, batch_first=False)

Bases: Module

forward(x)

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

training: bool

Module contents