MCPcopy
hub / github.com/shenweichen/DeepCTR-Torch / forward

Method forward

deepctr_torch/models/wdl.py:60–75  ·  view source on GitHub ↗
(self, X)

Source from the content-addressed store, hash-verified

58 self.to(device)
59
60 def forward(self, X):
61
62 sparse_embedding_list, dense_value_list = self.input_from_feature_columns(X, self.dnn_feature_columns,
63 self.embedding_dict)
64 logit = self.linear_model(X)
65
66 if self.use_dnn:
67 dnn_input = combined_dnn_input(sparse_embedding_list, dense_value_list)
68
69 dnn_output = self.dnn(dnn_input)
70 dnn_logit = self.dnn_linear(dnn_output)
71 logit += dnn_logit
72
73 y_pred = self.out(logit)
74
75 return y_pred

Callers

nothing calls this directly

Calls 2

combined_dnn_inputFunction · 0.85

Tested by

no test coverage detected