CEL

Code/train.py

@@ -142,7 +142,7 @@ def train(model, name, seed,gradient_accumulation_steps,mixup, threshold, lambda
 
 def cross_entropy(preds, target):
     log_preds = torch.log(preds)
-    loss = -torch.mean(target * log_preds)
+    loss = -torch.sum(target * log_preds) / preds.shape[0]
     return loss
 	
 def mixup_function(batch_of_matrices, batch_of_labels, l, t):