trading works!

NN/models/dqn_agent.py

@@ -221,7 +221,7 @@ class DQNAgent:
         # Check if mixed precision training should be used
         if torch.cuda.is_available() and hasattr(torch.cuda, 'amp') and 'DISABLE_MIXED_PRECISION' not in os.environ:
             self.use_mixed_precision = True
-            self.scaler = torch.amp.GradScaler('cuda')
+            self.scaler = torch.cuda.amp.GradScaler()
             logger.info("Mixed precision training enabled")
         else:
             self.use_mixed_precision = False
@@ -1083,6 +1083,11 @@ class DQNAgent:
             # Reset gradients
             self.optimizer.zero_grad()
             
+            # Ensure loss requires gradients before backward pass
+            if not total_loss.requires_grad:
+                logger.warning("Total loss tensor does not require gradients, skipping backward pass")
+                return 0.0
+            
             # Backward pass
             total_loss.backward()
             
@@ -1263,6 +1268,11 @@ class DQNAgent:
                         # Just use Q-value loss
                         loss = q_loss
                 
+                # Ensure loss requires gradients before backward pass
+                if not loss.requires_grad:
+                    logger.warning("Loss tensor does not require gradients, skipping backward pass")
+                    return 0.0
+                
                 # Backward pass with scaled gradients
                 self.scaler.scale(loss).backward()