print statics, save models and statics

scripts/ActorOnly+CEL/main_ActorOnly.py

@@ -50,24 +50,32 @@ for epoch in range(epochs):
     if epoch_rouge > best_rouge:
         best_rouge = epoch_rouge
         best_model_wts = copy.deepcopy(m.state_dict())
+        
+        # (over)write best model in file, and for Critic: save model_actor_only   
+        torch.save(m.state_dict(), 'model_actor_only_wts.pth')
+
+    time_elapsed = time.time() - since
+    print('Training + validation running already {:.0f}m {:.0f}s'.format(time_elapsed // 60, time_elapsed % 60))
+    print('Best val rouge so far: {:4f}'.format(best_rouge))
     
+    # save model weights after each validation
     torch.save(m.state_dict(), f'model_actor_only_wts{epoch}.pth')
+    
+   # (over)write val_rouge_history in file
+    torch.save(val_rouge_history, 'model_actor_only_val_rouge_hist.pt') 
+
 
 # after training completed
 time_elapsed = time.time() - since
 print('Training + validation complete in {:.0f}m {:.0f}s'.format(time_elapsed // 60, time_elapsed % 60))
 print('Best val rouge: {:4f}'.format(best_rouge))
 
-# write val_rouge_history in file
-pathlib.Path('hist').mkdir(parents=True, exist_ok=True)
-hist_path = pathlib.Path('hist')
-torch.save(val_rouge_history, hist_path/'model_actor_only_val_rouge_hist.pt') # all hist in one folder hist?
+
 
 # load best model weights
 m.load_state_dict(best_model_wts)
 
-# for Critic: save model_actor_only   
-torch.save(m.state_dict(), 'model_actor_only_wts.pth')
+
 # set sent_vecs 
 train_data.compute_sent_vecs(m, 'workspace/students/kreuzer/new_train')  #
 val_data.compute_sent_vecs(m, 'workspace/students/kreuzer/new_val')

scripts/ActorOnly+CEL/main_CrossEntropy.py

@@ -41,28 +41,38 @@ for epoch in range(epochs):
     val_rouge_history.append(epoch_rouge)    
     print('Validation Rouge Score: {:.4f}'.format(epoch_rouge))
 
+    
+
     # epoch completed, deep copy the best model sofar
     if epoch_rouge > best_rouge:
         best_rouge = epoch_rouge
         best_model_wts = copy.deepcopy(m.state_dict())
+
+        # (over)write best model in file, and for Critic: save model_actor_only   
+        torch.save(m.state_dict(), 'model_CE_wts.pth')
     
+    time_elapsed = time.time() - since
+    print('Training + validation running already {:.0f}m {:.0f}s'.format(time_elapsed // 60, time_elapsed % 60))
+    print('Best val rouge so far: {:4f}'.format(best_rouge))
+
+    # save model weights after each validation
     torch.save(m.state_dict(), f'model_CE_wts{epoch}.pth')
+    
+   # (over)write val_rouge_history in file
+    torch.save(val_rouge_history, 'model_CE_val_rouge_hist.pt') 
+
 
 # after training completed
 time_elapsed = time.time() - since
 print('Training + validation complete in {:.0f}m {:.0f}s'.format(time_elapsed // 60, time_elapsed % 60))
 print('Best val rouge: {:4f}'.format(best_rouge))
 
-# write val_rouge_history in file
-pathlib.Path('hist').mkdir(parents=True, exist_ok=True)
-hist_path = pathlib.Path('hist')
-torch.save(val_rouge_history, hist_path/'model_CE_val_rouge_hist.pt')
+
 
 # load best model weights
 m.load_state_dict(best_model_wts)
 
-#save best model
-torch.save(m.state_dict(), f'model_CE_wts.pth')
+
 
 # testing
 since = time.time()

scripts/Critic+ActorCritic/main_Critic.py

@@ -6,18 +6,18 @@ from models import *
 import time
 
 # loads dataset cnn_dailymail
-train_dataset = PreprocessedDataSet('/workspace/students/kreuzer/train')
-validation_dataset = PreprocessedDataSet('/workspace/students/kreuzer/val')
-test_dataset = PreprocessedDataSet('/workspace/students/kreuzer/test')
+train_dataset = PreprocessedDataSet(['/workspace/students/kreuzer/new_train'])
+validation_dataset = PreprocessedDataSet(['/workspace/students/kreuzer/new_val'])
+test_dataset = PreprocessedDataSet(['/workspace/students/kreuzer/new_test'])
 
 
 
 # hyperparameters
-epochs=20
+epochs=10 # <= actor trained epochs 
 batch_size=20
 learning_rate=0.001
 
-training_dataloader = torch.utils.data.DataLoader(train_dataset, batch_size=batch_size, shuffle=True, collate_fn = lambda x: x)
+train_dataloader = torch.utils.data.DataLoader(train_dataset, batch_size=batch_size, shuffle=True, collate_fn = lambda x: x)
 
 # load state dict from training
 model_actor_only = ActorOnlySummarisationModel()
@@ -39,11 +39,11 @@ for e in range(epochs):
     print('-' * 10)
 
     # train phase
-    epoch_loss = m.training_epoch(training_dataloader)  # collate_fn
+    epoch_loss = m.training_epoch(train_dataloader)  # collate_fn
     print('Train Loss: {:.4f}'.format(epoch_loss))
     
     # validation phase 
-    val_epoch_loss = m.test(dataset.validation)
+    val_epoch_loss = m.test(validation_dataset)
     val_loss_history.append(val_epoch_loss)    
     print('Validation Loss: {:.4f}'.format(val_epoch_loss))
 
@@ -51,6 +51,19 @@ for e in range(epochs):
     if val_epoch_loss < best_loss:
         best_loss = val_epoch_loss
         best_model_wts = copy.deepcopy(m.state_dict())
+        # (over)write best model in file
+        torch.save(m.state_dict(), 'model_critic_wts.pth')
+    
+    time_elapsed = time.time() - since
+    print('Training + validation running already {:.0f}m {:.0f}s'.format(time_elapsed // 60, time_elapsed % 60))
+    print('Best val result so far: {:4f}'.format(best_loss))
+
+    # save model weights after each validation
+    torch.save(m.state_dict(), f'model_critic_wts{epoch}.pth')
+    
+    # (over)write val_rouge_history in file
+    torch.save(val_loss_history, 'model_critic_val_loss_hist.pt') 
+
 
 # after training completed
 time_elapsed = time.time() - since
@@ -64,7 +77,7 @@ m.load_state_dict(best_model_wts)
 
 # testing
 since = time.time()
-test_loss = m.test(dataset.test)
+test_loss = m.test(test_dataset)
 print('Test rouge score difference: {:.4f}'.format(test_loss))
 
 # after testing completed