Change BCELoss in MLP model (microsoft#756)

Author: cuicorey

qlib/contrib/model/pytorch_nn.py

@@ -267,7 +267,7 @@ def get_loss(self, pred, w, target, loss_type):
             loss = torch.mul(sqr_loss, w).mean()
             return loss
         elif loss_type == "binary":
-            loss = nn.BCELoss(weight=w)
+            loss = nn.BCEWithLogitsLoss(weight=w)
             return loss(pred, target)
         else:
             raise NotImplementedError("loss {} is not supported!".format(loss_type))
@@ -334,16 +334,8 @@ def __init__(self, input_dim, output_dim, layers=(256, 512, 768, 512, 256, 128,
             dnn_layers.append(seq)
         drop_input = nn.Dropout(0.05)
         dnn_layers.append(drop_input)
-        if loss == "mse":
-            fc = nn.Linear(hidden_units, output_dim)
-            dnn_layers.append(fc)
-
-        elif loss == "binary":
-            fc = nn.Linear(hidden_units, output_dim)
-            sigmoid = nn.Sigmoid()
-            dnn_layers.append(nn.Sequential(fc, sigmoid))
-        else:
-            raise NotImplementedError("loss {} is not supported!".format(loss))
+        fc = nn.Linear(hidden_units, output_dim)
+        dnn_layers.append(fc)
         # optimizer
         self.dnn_layers = nn.ModuleList(dnn_layers)
         self._weight_init()