Update modeling_gptbert.py

modeling_gptbert.py

@@ -864,6 +864,10 @@ class Classifier(nn.Module):
         # self.emb2vocab = CastedLinear(config.hidden_size, num_labels, bias=True)
         self.dropout = nn.Dropout(drop_out)
 
+        print(f"Hidden Size: {config.hidden_size}")
+        print(f"Intermediate Size: {config.intermediate_size}")
+        print(f"Proj_std: {math.sqrt(2.0 / (hidden_size + intermediate_size))}")
+
         self.initialize(config.hidden_size, config.intermediate_size, num_labels)
 
     @torch.no_grad()