Merge pull request 'feat/checkpoints' (#14) from feat/checkpoints into develop

Reviewed-on: #14

README.md

@@ -26,15 +26,22 @@ pip install git+https://gitea.fabelous.app/Machine-Learning/aiuNN.git
 Here's a basic example of how to use `aiuNN` for image upscaling:
 
 ```python src/main.py
-from aiia import AIIABase
+from aiia import AIIABase, AIIAConfig
 from aiunn import aiuNN, aiuNNTrainer
 import pandas as pd
 from torchvision import transforms
 
+# Create a configuration and build a base model.
+config = AIIAConfig()
+ai_config = aiuNNConfig()
+
+base_model = AIIABase(config)
+upscaler = aiuNN(config=ai_config)
+
 # Load your base model and upscaler
 pretrained_model_path = "path/to/aiia/model"
-base_model = AIIABase.load(pretrained_model_path, precision="bf16")
-upscaler = aiuNN(base_model)
+base_model = AIIABase.from_pretrained(pretrained_model_path)
+upscaler.load_base_model(base_model)
 
 # Create trainer with your dataset class
 trainer = aiuNNTrainer(upscaler, dataset_class=UpscaleDataset)

setup.py

@@ -2,7 +2,7 @@ from setuptools import setup, find_packages
 
 setup(
     name="aiunn",
-    version="0.2.0",
+    version="0.2.1",
     packages=find_packages(where="src"),
     package_dir={"": "src"},
     install_requires=[

src/aiunn/__init__.py

@@ -3,4 +3,4 @@ from .upsampler.aiunn import aiuNN
 from .upsampler.config import aiuNNConfig
 from .inference.inference import aiuNNInference
 
-__version__ = "0.2.0"
+__version__ = "0.2.1"

src/aiunn/finetune/trainer.py

@@ -10,6 +10,7 @@ from torch.utils.checkpoint import checkpoint
 import gc
 import time
 import shutil
+import datetime
 
 
 class EarlyStopping:
@@ -50,10 +51,16 @@ class aiuNNTrainer:
         self.optimizer = None
         self.scaler = GradScaler()
         self.best_loss = float('inf')
-        self.use_checkpointing = True
+        self.csv_path = None
+        self.checkpoint_dir = None
         self.data_loader = None
         self.validation_loader = None
-        self.log_dir = None
+        self.last_checkpoint_time = time.time()
+        self.checkpoint_interval = 2 * 60 * 60  # 2 hours
+        self.last_22_date = None
+        self.recent_checkpoints = []
+        self.current_epoch = 0
+
         
     def load_data(self, dataset_params=None, batch_size=1, validation_split=0.2, custom_train_dataset=None, custom_val_dataset=None):
         """
@@ -110,23 +117,19 @@ class aiuNNTrainer:
         return self.data_loader, self.validation_loader
     
     def _setup_logging(self, output_path):
-        """Set up directory structure for logging and model checkpoints"""
-        timestamp = time.strftime("%Y%m%d-%H%M%S")
-        self.log_dir = os.path.join(output_path, f"training_run_{timestamp}")
-        os.makedirs(self.log_dir, exist_ok=True)
-        
+        """Set up basic logging and checkpoint directory"""
         # Create checkpoint directory
-        self.checkpoint_dir = os.path.join(self.log_dir, "checkpoints")
+        self.checkpoint_dir = os.path.join(output_path, "checkpoints")
         os.makedirs(self.checkpoint_dir, exist_ok=True)
         
         # Set up CSV logging
-        self.csv_path = os.path.join(self.log_dir, 'training_log.csv')
+        self.csv_path = os.path.join(output_path, 'training_log.csv')
         with open(self.csv_path, mode='w', newline='') as file:
             writer = csv.writer(file)
             if self.validation_loader:
-                writer.writerow(['Epoch', 'Train Loss', 'Validation Loss', 'Improved'])
+                writer.writerow(['Epoch', 'Train Loss', 'Validation Loss'])
             else:
-                writer.writerow(['Epoch', 'Train Loss', 'Improved'])
+                writer.writerow(['Epoch', 'Train Loss'])
     
     def _evaluate(self):
         """Evaluate the model on validation data"""
@@ -152,63 +155,99 @@ class aiuNNTrainer:
         self.model.train()
         return val_loss
     
-    def _save_checkpoint(self, epoch, is_best=False):
-        """Save model checkpoint"""
-        checkpoint_path = os.path.join(self.checkpoint_dir, f"epoch_{epoch}.pt")
-        best_model_path = os.path.join(self.log_dir, "best_model")
+    def _save_checkpoint(self, epoch, batch_count, is_best=False, is_22=False):
+        """Save checkpoint with support for regular, best, and 22:00 saves"""
+        if is_22:
+            today = datetime.datetime.now(datetime.timezone(datetime.timedelta(hours=2))).date()
+            checkpoint_name = f"checkpoint_22h_{today.strftime('%Y%m%d')}.pt"
+        else:
+            checkpoint_name = f"checkpoint_epoch{epoch}_batch{batch_count}.pt"
             
-        # Save the model checkpoint
-        self.model.save(checkpoint_path)
+        checkpoint_path = os.path.join(self.checkpoint_dir, checkpoint_name)
         
-        # If this is the best model so far, copy it to best_model
+        checkpoint_data = {
+            'epoch': epoch,
+            'batch': batch_count,
+            'model_state_dict': self.model.state_dict(),
+            'optimizer_state_dict': self.optimizer.state_dict(),
+            'best_loss': self.best_loss,
+            'scaler_state_dict': self.scaler.state_dict()
+        }
+        
+        torch.save(checkpoint_data, checkpoint_path)
+        
+        # Save best model separately
         if is_best:
-            if os.path.exists(best_model_path):
-                shutil.rmtree(best_model_path)
-            self.model.save(best_model_path)
-            print(f"Saved new best model with loss: {self.best_loss:.6f}")
+            best_model_path = os.path.join(os.path.dirname(self.checkpoint_dir), "best_model")
+            self.model.save_pretrained(best_model_path)
+            
+        return checkpoint_path
+
+    def _handle_checkpoints(self, epoch, batch_count, is_improved):
+        """Handle periodic and 22:00 checkpoint saving"""
+        current_time = time.time()
+        current_dt = datetime.datetime.now(datetime.timezone(datetime.timedelta(hours=2)))
+        
+        # Regular interval checkpoint
+        if (current_time - self.last_checkpoint_time) >= self.checkpoint_interval:
+            self._save_checkpoint(epoch, batch_count, is_improved)
+            self.last_checkpoint_time = current_time
+
+        # Special 22:00 checkpoint
+        is_22_oclock = current_dt.hour == 22 and current_dt.minute < 15
+        if is_22_oclock and self.last_22_date != current_dt.date():
+            self._save_checkpoint(epoch, batch_count, is_improved, is_22=True)
+            self.last_22_date = current_dt.date()
 
     def finetune(self, output_path, epochs=10, lr=1e-4, patience=3, min_delta=0.001):
-        """
-        Finetune the upscaler model
-        
-        Args:
-            output_path (str): Directory to save models and logs
-            epochs (int): Maximum number of training epochs
-            lr (float): Learning rate
-            patience (int): Early stopping patience
-            min_delta (float): Minimum improvement for early stopping
-        """
-        # Check if data is loaded
+        """Finetune the upscaler model"""
         if self.data_loader is None:
             raise ValueError("Data not loaded. Call load_data first.")
         
-        # Setup optimizer
+        # Setup optimizer and directories
         self.optimizer = optim.Adam(self.model.parameters(), lr=lr)
+        self.checkpoint_dir = os.path.join(output_path, "checkpoints")
+        os.makedirs(self.checkpoint_dir, exist_ok=True)
         
-        # Set up logging
-        self._setup_logging(output_path)
+        # Setup CSV logging
+        self.csv_path = os.path.join(output_path, 'training_log.csv')
+        with open(self.csv_path, mode='w', newline='') as file:
+            writer = csv.writer(file)
+            header = ['Epoch', 'Train Loss', 'Validation Loss'] if self.validation_loader else ['Epoch', 'Train Loss']
+            writer.writerow(header)
+        
+        # Load existing checkpoint if available
+        checkpoint_info = self.load_checkpoint()
+        start_epoch = checkpoint_info[0] if checkpoint_info else 0
+        start_batch = checkpoint_info[1] if checkpoint_info else 0
         
         # Setup early stopping
         early_stopping = EarlyStopping(patience=patience, min_delta=min_delta)
+        self.best_loss = float('inf')
         
         # Training loop
         self.model.train()
-        
-        for epoch in range(epochs):
-            # Training phase
+        for epoch in range(start_epoch, epochs):
+            self.current_epoch = epoch
             epoch_loss = 0.0
-            progress_bar = tqdm(self.data_loader, desc=f"Epoch {epoch + 1}/{epochs}")
             
-            for low_res, high_res in progress_bar:
-                # Move data to GPU with channels_last format where possible
+            train_batches = list(enumerate(self.data_loader))
+            start_idx = start_batch if epoch == start_epoch else 0
+            
+            progress_bar = tqdm(train_batches[start_idx:], 
+                            initial=start_idx, 
+                            total=len(train_batches),
+                            desc=f"Epoch {epoch + 1}/{epochs}")
+
+            for batch_idx, (low_res, high_res) in progress_bar:
+                # Training step
                 low_res = low_res.to(self.device, non_blocking=True).to(memory_format=torch.channels_last)
                 high_res = high_res.to(self.device, non_blocking=True)
                 
                 self.optimizer.zero_grad()
                 
                 with autocast(device_type=self.device.type):
-                    if self.use_checkpointing:
-                        # Ensure the input tensor requires gradient so that checkpointing records the computation graph
+                    if hasattr(self, 'use_checkpointing') and self.use_checkpointing:
                         low_res.requires_grad_()
                         outputs = checkpoint(self.model, low_res)
                     else:
@@ -222,69 +261,109 @@ class aiuNNTrainer:
                 epoch_loss += loss.item()
                 progress_bar.set_postfix({'loss': loss.item()})
                 
-                # Optionally delete variables to free memory
+                # Handle checkpoints
+                self._handle_checkpoints(epoch + 1, batch_idx + 1, loss.item() < self.best_loss)
+                
                 del low_res, high_res, outputs, loss
             
-            # Calculate average epoch loss
+            # End of epoch processing
             avg_train_loss = epoch_loss / len(self.data_loader)
             
-            # Validation phase (if validation loader exists)
+            # Validation phase
             if self.validation_loader:
                 val_loss = self._evaluate() / len(self.validation_loader)
                 is_improved = val_loss < self.best_loss
                 if is_improved:
                     self.best_loss = val_loss
                 
-                # Log results
-                print(f"Epoch {epoch + 1}/{epochs}, Train Loss: {avg_train_loss:.6f}, Val Loss: {val_loss:.6f}")
+                # Log to CSV
                 with open(self.csv_path, mode='a', newline='') as file:
                     writer = csv.writer(file)
-                    writer.writerow([epoch + 1, avg_train_loss, val_loss, "Yes" if is_improved else "No"])
+                    writer.writerow([epoch + 1, avg_train_loss, val_loss])
+                    
+                print(f"Epoch {epoch + 1}/{epochs}, Train Loss: {avg_train_loss:.6f}, Val Loss: {val_loss:.6f}")
             else:
-                # If no validation, use training loss for improvement tracking
                 is_improved = avg_train_loss < self.best_loss
                 if is_improved:
                     self.best_loss = avg_train_loss
                 
-                # Log results
-                print(f"Epoch {epoch + 1}/{epochs}, Train Loss: {avg_train_loss:.6f}")
+                # Log to CSV
                 with open(self.csv_path, mode='a', newline='') as file:
                     writer = csv.writer(file)
-                    writer.writerow([epoch + 1, avg_train_loss, "Yes" if is_improved else "No"])
+                    writer.writerow([epoch + 1, avg_train_loss])
                     
-            # Save checkpoint
-            self._save_checkpoint(epoch + 1, is_best=is_improved)
+                print(f"Epoch {epoch + 1}/{epochs}, Train Loss: {avg_train_loss:.6f}")
             
-            # Perform garbage collection and clear GPU cache after each epoch
-            gc.collect()
-            torch.cuda.empty_cache()
+            # Save best model if improved
+            if is_improved:
+                best_model_path = os.path.join(output_path, "best_model")
+                self.model.save_pretrained(best_model_path)
             
             # Check early stopping
-            early_stopping(val_loss if self.validation_loader else avg_train_loss)
-            if early_stopping.early_stop:
+            if early_stopping(val_loss if self.validation_loader else avg_train_loss):
                 print(f"Early stopping triggered at epoch {epoch + 1}")
                 break
             
+            # Cleanup
+            gc.collect()
+            torch.cuda.empty_cache()
+        
         return self.best_loss
 
+    def load_checkpoint(self, specific_checkpoint=None):
+        """Enhanced checkpoint loading with specific checkpoint support"""
+        if specific_checkpoint:
+            checkpoint_path = os.path.join(self.checkpoint_dir, specific_checkpoint)
+        else:
+            checkpoint_files = [f for f in os.listdir(self.checkpoint_dir) 
+                              if f.startswith("checkpoint_") and f.endswith(".pt")]
+            if not checkpoint_files:
+                return None
+            
+            checkpoint_files.sort(key=lambda x: os.path.getmtime(
+                os.path.join(self.checkpoint_dir, x)), reverse=True)
+            checkpoint_path = os.path.join(self.checkpoint_dir, checkpoint_files[0])
+
+        if not os.path.exists(checkpoint_path):
+            return None
+
+        checkpoint = torch.load(checkpoint_path, map_location=self.device)
+        self.model.load_state_dict(checkpoint['model_state_dict'])
+        self.optimizer.load_state_dict(checkpoint['optimizer_state_dict'])
+        self.scaler.load_state_dict(checkpoint['scaler_state_dict'])
+        self.best_loss = checkpoint['best_loss']
+
+        print(f"Loaded checkpoint from {checkpoint_path}")
+        return checkpoint['epoch'], checkpoint['batch']
+                
     def save(self, output_path=None):
         """
         Save the best model to the specified path
         
         Args:
-            output_path (str, optional): Path to save the model. If None, uses the best model from training.
+            output_path (str, optional): Path to save the model. If None, tries to use the checkpoint directory from training.
+        
+        Returns:
+            str: Path where the model was saved
+        
+        Raises:
+            ValueError: If no output path is specified and no checkpoint directory exists
         """
-        if output_path is None and self.log_dir is not None:
-            best_model_path = os.path.join(self.log_dir, "best_model")
+        if output_path is None and self.checkpoint_dir is not None:
+            # First try to copy the best model if it exists
+            best_model_path = os.path.join(os.path.dirname(self.checkpoint_dir), "best_model")
             if os.path.exists(best_model_path):
-                print(f"Best model already saved at {best_model_path}")
-                return best_model_path
+                output_path = os.path.join(os.path.dirname(self.checkpoint_dir), "final_model")
+                shutil.copytree(best_model_path, output_path, dirs_exist_ok=True)
+                print(f"Copied best model to {output_path}")
+                return output_path
             else:
-                output_path = os.path.join(self.log_dir, "final_model")
+                # If no best model exists, save current model state
+                output_path = os.path.join(os.path.dirname(self.checkpoint_dir), "final_model")
         
         if output_path is None:
-            raise ValueError("No output path specified and no training has been done yet.")
+            raise ValueError("No output path specified and no checkpoint directory exists from training.")
         
-        self.model.save(output_path)
+        self.model.save_pretrained(output_path)
         print(f"Model saved to {output_path}")
         return output_path