chore: readable print logs & separate clearml graphs

trainingModel/Training.py

@@ -130,7 +130,7 @@ def train_model(
         epoch_accuracies[epoch] = epoch_acc_avg
 
         print(f"\n--- Epoch {epoch + 1}: ---")
-        print(f'Train loss={epoch_loss_avg:.4f}\nTrain accuracy={epoch_acc_avg:.4f}')
+        print(f'Train loss={epoch_loss_avg:.4f}\nTrain accuracy={epoch_acc_avg:.4f}\n')
 
         # ----------------------
         # validation loop
@@ -156,7 +156,7 @@ def train_model(
         current_val_accuracy = val_accuracy_fn.compute().item()
         val_accuracies[epoch] = current_val_accuracy
 
-        print(f"Epoch {epoch+1}: val acc={current_val_accuracy:.4f}")
+        print(f"\nEpoch {epoch+1}: val acc={current_val_accuracy:.4f}")
 
         # keep track of best validation accuracy and save best model so far
         if current_val_accuracy > best_accuracy:
@@ -164,7 +164,7 @@ def train_model(
             torch.save(model.state_dict(), save_path)
         
 
-        print(f'Epoch {epoch + 1} validation complete')
+        print(f'Epoch {epoch + 1} validation complete\n')
 
     print(f"\nTraining finished. Best val accuracy: {best_accuracy:.4f}")
     print(f"Best model weights saved to: {save_path}")

trainingModel/run_training.py

@@ -140,21 +140,20 @@ training_metrics = train_model(
 # ----------- Log metrics to ClearML -----------
 # Per-batch training losses and accuracies
 for i, loss in enumerate(training_metrics["batch_losses"]):
-    training_logger.report_scalar("train_batch", "loss", value=loss, iteration=i)
+    training_logger.report_scalar("training batch loss", "loss", value=loss, iteration=i)
 
 for i, acc in enumerate(training_metrics["batch_accuracies"]):
-    training_logger.report_scalar("train_batch", "accuracy", value=acc, iteration=i)
-
+    training_logger.report_scalar("training batch accuracy", "accuracy", value=acc, iteration=i)
 
 # Per-epoch training losses and accuracies
 epoch_metrics = zip(training_metrics["epoch_losses"], training_metrics["epoch_accuracies"])
 for epoch, (loss, acc) in enumerate(epoch_metrics):
-    training_logger.report_scalar("train_epoch", "loss", loss, iteration=epoch)
-    training_logger.report_scalar("train_epoch", "accuracy", acc, iteration=epoch)
+    training_logger.report_scalar("training epoch loss", "loss", loss, iteration=epoch)
+    training_logger.report_scalar("training epoch accuracy", "accuracy", acc, iteration=epoch)
 
 # Per-epoch validation accuracies
 for epoch, acc in enumerate(training_metrics["val_accuracies"]):
-    training_logger.report_scalar("validation_epoch", "accuracy", value=acc, iteration=epoch)
+    training_logger.report_scalar("validation epoch accuracy", "accuracy", value=acc, iteration=epoch)
 
 training_logger.report_single_value("best_val_accuracy", training_metrics["best_accuracy"])