added result, 18.5276

image-inpainting/src/architecture.py

@@ -18,6 +18,20 @@ def init_weights(m):
     elif isinstance(m, nn.BatchNorm2d):
         nn.init.constant_(m.weight, 1)
         nn.init.constant_(m.bias, 0)
+    elif isinstance(m, nn.GroupNorm):
+        if m.weight is not None:
+            nn.init.constant_(m.weight, 1)
+        if m.bias is not None:
+            nn.init.constant_(m.bias, 0)
+
+
+def _make_norm(num_channels: int) -> nn.Module:
+    """Batch-size independent normalization (works well for batch_size=1 eval)."""
+    # Choose a group count that divides num_channels.
+    num_groups = min(32, num_channels)
+    while num_groups > 1 and (num_channels % num_groups) != 0:
+        num_groups //= 2
+    return nn.GroupNorm(num_groups=num_groups, num_channels=num_channels)
 
 
 class ChannelAttention(nn.Module):
@@ -73,7 +87,7 @@ class ConvBlock(nn.Module):
     def __init__(self, in_channels, out_channels, kernel_size=3, padding=1, dropout=0.0):
         super().__init__()
         self.conv = nn.Conv2d(in_channels, out_channels, kernel_size, padding=padding)
-        self.bn = nn.BatchNorm2d(out_channels)
+        self.bn = _make_norm(out_channels)
         self.relu = nn.LeakyReLU(0.1, inplace=True)
         self.dropout = nn.Dropout2d(dropout) if dropout > 0 else nn.Identity()
     
@@ -85,9 +99,9 @@ class ResidualConvBlock(nn.Module):
     def __init__(self, channels, dropout=0.0):
         super().__init__()
         self.conv1 = nn.Conv2d(channels, channels, 3, padding=1)
-        self.bn1 = nn.BatchNorm2d(channels)
+        self.bn1 = _make_norm(channels)
         self.conv2 = nn.Conv2d(channels, channels, 3, padding=1)
-        self.bn2 = nn.BatchNorm2d(channels)
+        self.bn2 = _make_norm(channels)
         self.relu = nn.LeakyReLU(0.1, inplace=True)
         self.dropout = nn.Dropout2d(dropout) if dropout > 0 else nn.Identity()
     
@@ -100,24 +114,24 @@ class ResidualConvBlock(nn.Module):
         return self.relu(out)
 
 
-class DilatedResidualBlock(nn.Module):
+class GatedConvBlock(nn.Module):
-    """Residual block with dilated convolutions for larger receptive field"""
+    """Gated convolution block (helps the network condition on the mask channel)."""
-    def __init__(self, channels, dilation=2, dropout=0.0):
+    def __init__(self, in_channels, out_channels, kernel_size=3, padding=1, dropout=0.0):
         super().__init__()
-        self.conv1 = nn.Conv2d(channels, channels, 3, padding=dilation, dilation=dilation)
+        self.feature = nn.Conv2d(in_channels, out_channels, kernel_size, padding=padding)
-        self.bn1 = nn.BatchNorm2d(channels)
+        self.gate = nn.Conv2d(in_channels, out_channels, kernel_size, padding=padding)
-        self.conv2 = nn.Conv2d(channels, channels, 3, padding=dilation, dilation=dilation)
+        self.norm = _make_norm(out_channels)
-        self.bn2 = nn.BatchNorm2d(channels)
+        self.act = nn.LeakyReLU(0.1, inplace=True)
-        self.relu = nn.LeakyReLU(0.1, inplace=True)
         self.dropout = nn.Dropout2d(dropout) if dropout > 0 else nn.Identity()
 
     def forward(self, x):
-        residual = x
+        feat = self.feature(x)
-        out = self.relu(self.bn1(self.conv1(x)))
+        gate = torch.sigmoid(self.gate(x))
+        out = feat * gate
+        out = self.norm(out)
+        out = self.act(out)
         out = self.dropout(out)
-        out = self.bn2(self.conv2(out))
+        return out
-        out = out + residual
-        return self.relu(out)
 
 
 class DownBlock(nn.Module):
@@ -167,7 +181,7 @@ class MyModel(nn.Module):
         
         # Initial convolution with larger receptive field
         self.init_conv = nn.Sequential(
-            ConvBlock(n_in_channels, base_channels, kernel_size=7, padding=3),
+            GatedConvBlock(n_in_channels, base_channels, kernel_size=7, padding=3, dropout=dropout),
             ConvBlock(base_channels, base_channels),
             ResidualConvBlock(base_channels)
         )
@@ -178,12 +192,11 @@ class MyModel(nn.Module):
         self.down3 = DownBlock(base_channels * 4, base_channels * 8, dropout=dropout)
         self.down4 = DownBlock(base_channels * 8, base_channels * 16, dropout=dropout)
         
-        # Bottleneck with multi-scale dilated convolutions (ASPP-style)
+        # Bottleneck with multiple residual blocks
         self.bottleneck = nn.Sequential(
             ConvBlock(base_channels * 16, base_channels * 16, dropout=dropout),
             ResidualConvBlock(base_channels * 16, dropout=dropout),
-            DilatedResidualBlock(base_channels * 16, dilation=2, dropout=dropout),
+            ResidualConvBlock(base_channels * 16, dropout=dropout),
-            DilatedResidualBlock(base_channels * 16, dilation=4, dropout=dropout),
             ResidualConvBlock(base_channels * 16, dropout=dropout),
             CBAM(base_channels * 16)
         )

image-inpainting/src/datasets.py

@@ -10,7 +10,7 @@ import numpy as np
 import random
 import glob
 import os
-from PIL import Image, ImageEnhance
+from PIL import Image
 
 IMAGE_DIMENSION = 100
 
@@ -32,75 +32,54 @@ def resize(img: Image):
         transforms.CenterCrop((IMAGE_DIMENSION, IMAGE_DIMENSION))
     ])
     return resize_transforms(img)
+
+
+def augment_geometric(img: Image.Image) -> Image.Image:
+    """Lightweight, label-preserving augmentation (safe for train/val/test splits)."""
+    # Horizontal flip
+    if random.random() < 0.5:
+        img = img.transpose(Image.Transpose.FLIP_LEFT_RIGHT)
+    # Vertical flip (less frequent)
+    if random.random() < 0.2:
+        img = img.transpose(Image.Transpose.FLIP_TOP_BOTTOM)
+    # 90-degree rotations (no interpolation artifacts)
+    r = random.random()
+    if r < 0.25:
+        img = img.transpose(Image.Transpose.ROTATE_90)
+    elif r < 0.5:
+        img = img.transpose(Image.Transpose.ROTATE_180)
+    elif r < 0.75:
+        img = img.transpose(Image.Transpose.ROTATE_270)
+    return img
 def preprocess(input_array: np.ndarray):
     input_array = np.asarray(input_array, dtype=np.float32) / 255.0
     return input_array
 
 class ImageDataset(torch.utils.data.Dataset):
     """
-    Dataset class for loading images from a folder with data augmentation
+    Dataset class for loading images from a folder
     """
 
-    def __init__(self, datafolder: str, augment: bool = True):
+    def __init__(self, datafolder: str):
         self.imagefiles = sorted(glob.glob(os.path.join(datafolder,"**","*.jpg"),recursive=True))
-        self.augment = augment
 
     def __len__(self):
         return len(self.imagefiles)
         
-    def augment_image(self, image: Image) -> Image:
-        """Apply random augmentations to image"""
-        # Random horizontal flip
-        if random.random() > 0.5:
-            image = image.transpose(Image.FLIP_LEFT_RIGHT)
-        
-        # Random vertical flip
-        if random.random() > 0.5:
-            image = image.transpose(Image.FLIP_TOP_BOTTOM)
-        
-        # Random rotation (90, 180, 270 degrees)
-        if random.random() > 0.5:
-            angle = random.choice([90, 180, 270])
-            image = image.rotate(angle)
-        
-        # Random brightness adjustment
-        if random.random() > 0.5:
-            enhancer = ImageEnhance.Brightness(image)
-            factor = random.uniform(0.8, 1.2)
-            image = enhancer.enhance(factor)
-        
-        # Random contrast adjustment
-        if random.random() > 0.5:
-            enhancer = ImageEnhance.Contrast(image)
-            factor = random.uniform(0.8, 1.2)
-            image = enhancer.enhance(factor)
-        
-        # Random color adjustment
-        if random.random() > 0.5:
-            enhancer = ImageEnhance.Color(image)
-            factor = random.uniform(0.8, 1.2)
-            image = enhancer.enhance(factor)
-        
-        return image
-    
     def __getitem__(self, idx:int):
         index = int(idx)
         
-        image = Image.open(self.imagefiles[index])
+        image = Image.open(self.imagefiles[index]).convert("RGB")
-        image = resize(image)
+        image = augment_geometric(image)
-        
+        image = np.asarray(resize(image))
-        # Apply augmentation if enabled
-        if self.augment:
-            image = self.augment_image(image)
-        
-        image = np.asarray(image)
         image = preprocess(image)
 
-        # Vary spacing and offset more for additional diversity
+        # Sample a grid-mask similar in density to the challenge testset (~8% known pixels).
-        spacing_x = random.randint(2,7)
+        # IMPORTANT: offset ranges must be tied to spacing to avoid accidental distribution shift.
-        spacing_y = random.randint(2,7)
+        spacing_x = random.randint(4, 6)
-        offset_x = random.randint(0,10)
+        spacing_y = random.randint(2, 4)
-        offset_y = random.randint(0,10)
+        offset_x = random.randint(0, spacing_x - 1)
+        offset_y = random.randint(0, spacing_y - 1)
         spacing = (spacing_x, spacing_y)
         offset = (offset_x, offset_y)
         input_array, known_array = create_arrays_from_image(image.copy(), offset, spacing)

image-inpainting/src/main.py

@@ -24,22 +24,22 @@ if __name__ == '__main__':
     config_dict['results_path'] = os.path.join(project_root, "results")
     config_dict['data_path'] = os.path.join(project_root, "data", "dataset")
     config_dict['device'] = None
-    config_dict['learningrate'] = 2e-4  # Slightly lower for more stable training
+    config_dict['learningrate'] = 3e-4  # Optimal learning rate for AdamW
-    config_dict['weight_decay'] = 5e-5  # Reduced for less aggressive regularization
+    config_dict['weight_decay'] = 1e-4  # Slightly higher for better regularization
-    config_dict['n_updates'] = 8000  # More updates for better convergence
+    config_dict['n_updates'] = 5000  # More updates for better convergence
-    config_dict['batchsize'] = 12  # Larger batch for more stable gradients
+    config_dict['batchsize'] = 8  # Smaller batch for better gradient estimates
-    config_dict['early_stopping_patience'] = 15  # More patience for complex model
+    config_dict['early_stopping_patience'] = 10  # More patience for complex model
     config_dict['use_wandb'] = False
 
     config_dict['print_train_stats_at'] = 10
     config_dict['print_stats_at'] = 100
-    config_dict['plot_at'] = 400
+    config_dict['plot_at'] = 300
-    config_dict['validate_at'] = 200  # Validate frequently but not too often
+    config_dict['validate_at'] = 300  # Validate more frequently
 
     network_config = {
         'n_in_channels': 4,
-        'base_channels': 32,  # Smaller base for efficiency, depth compensates
+        'base_channels': 48,  # Good balance between capacity and memory
-        'dropout': 0.15  # Slightly more regularization with augmentation
+        'dropout': 0.1  # Regularization
     }
     
     config_dict['network_config'] = network_config

image-inpainting/src/train.py

@@ -21,8 +21,8 @@ import wandb
 
 
 class CombinedLoss(nn.Module):
-    """Combined loss: MSE + L1 + Edge-aware component for better reconstruction"""
+    """Combined loss: MSE + L1 + SSIM-like perceptual component"""
-    def __init__(self, mse_weight=0.7, l1_weight=0.8, edge_weight=0.2):
+    def __init__(self, mse_weight=1.0, l1_weight=0.5, edge_weight=0.1):
         super().__init__()
         self.mse_weight = mse_weight
         self.l1_weight = l1_weight
@@ -84,24 +84,20 @@ def train(seed, testset_ratio, validset_ratio, data_path, results_path, early_st
     plotpath = os.path.join(results_path, "plots")
     os.makedirs(plotpath, exist_ok=True)
 
-    # Create dataset with augmentation for training, without for validation/test
+    image_dataset = datasets.ImageDataset(datafolder=data_path)
-    image_dataset_full = datasets.ImageDataset(datafolder=data_path, augment=False)
 
-    n_total = len(image_dataset_full)
+    n_total = len(image_dataset)
     n_test = int(n_total * testset_ratio)
     n_valid = int(n_total * validset_ratio)
     n_train = n_total - n_test - n_valid
     indices = np.random.permutation(n_total)
+    dataset_train = Subset(image_dataset, indices=indices[0:n_train])
+    dataset_valid = Subset(image_dataset, indices=indices[n_train:n_train + n_valid])
+    dataset_test = Subset(image_dataset, indices=indices[n_train + n_valid:n_total])
 
-    # Create augmented dataset for training
+    assert len(image_dataset) == len(dataset_train) + len(dataset_test) + len(dataset_valid)
-    image_dataset_train = datasets.ImageDataset(datafolder=data_path, augment=True)
-    dataset_train = Subset(image_dataset_train, indices=indices[0:n_train])
-    dataset_valid = Subset(image_dataset_full, indices=indices[n_train:n_train + n_valid])
-    dataset_test = Subset(image_dataset_full, indices=indices[n_train + n_valid:n_total])
 
-    assert n_total == len(dataset_train) + len(dataset_test) + len(dataset_valid)
+    del image_dataset
-
-    del image_dataset_full, image_dataset_train
 
     dataloader_train = DataLoader(dataset=dataset_train, batch_size=batchsize,
                                   num_workers=0, shuffle=True)
@@ -115,19 +111,15 @@ def train(seed, testset_ratio, validset_ratio, data_path, results_path, early_st
     network.to(device)
     network.train()
 
-    # defining the loss - combined loss with optimized weights
+    # defining the loss - combined loss for better reconstruction
-    combined_loss = CombinedLoss(mse_weight=0.7, l1_weight=0.8, edge_weight=0.2).to(device)
+    combined_loss = CombinedLoss(mse_weight=1.0, l1_weight=0.5, edge_weight=0.1).to(device)
     mse_loss = torch.nn.MSELoss()  # Keep for evaluation
 
     # defining the optimizer with AdamW for better weight decay handling
     optimizer = torch.optim.AdamW(network.parameters(), lr=learningrate, weight_decay=weight_decay)
     
-    # Learning rate scheduler with better configuration
+    # Learning rate scheduler for better convergence
-    scheduler = CosineAnnealingWarmRestarts(optimizer, T_0=100, T_mult=2, eta_min=1e-7)
+    scheduler = CosineAnnealingWarmRestarts(optimizer, T_0=50, T_mult=2, eta_min=1e-6)
-    
-    # Mixed precision training for faster computation and lower memory usage
-    scaler = torch.cuda.amp.GradScaler() if device.type == 'cuda' else None
-    use_amp = scaler is not None
 
     if use_wandb:
         wandb.watch(network, mse_loss, log="all", log_freq=10)
@@ -136,13 +128,10 @@ def train(seed, testset_ratio, validset_ratio, data_path, results_path, early_st
     counter = 0
     best_validation_loss = np.inf
     loss_list = []
-    accumulation_steps = 2  # Gradient accumulation for effective larger batch size
 
     saved_model_path = os.path.join(results_path, "best_model.pt")
 
     print(f"Started training on device {device}")
-    print(f"Using mixed precision: {use_amp}")
-    print(f"Gradient accumulation steps: {accumulation_steps}")
 
     while i < n_updates:
 
@@ -153,33 +142,21 @@ def train(seed, testset_ratio, validset_ratio, data_path, results_path, early_st
             if (i + 1) % print_train_stats_at == 0:
                 print(f'Update Step {i + 1} of {n_updates}: Current loss: {loss_list[-1]}')
 
-            # Use mixed precision if available
+            optimizer.zero_grad()
-            if use_amp:
-                with torch.cuda.amp.autocast():
-                    output = network(input)
-                    loss = combined_loss(output, target)
-                loss = loss / accumulation_steps
-                scaler.scale(loss).backward()
-            else:
-                output = network(input)
-                loss = combined_loss(output, target)
-                loss = loss / accumulation_steps
-                loss.backward()
 
-            # Gradient accumulation - update weights every accumulation_steps
+            output = network(input)
-            if (i + 1) % accumulation_steps == 0:
-                if use_amp:
-                    scaler.unscale_(optimizer)
-                    torch.nn.utils.clip_grad_norm_(network.parameters(), max_norm=1.0)
-                    scaler.step(optimizer)
-                    scaler.update()
-                else:
-                    torch.nn.utils.clip_grad_norm_(network.parameters(), max_norm=1.0)
-                    optimizer.step()
-                optimizer.zero_grad()
-                scheduler.step(i / n_updates)
 
-            loss_list.append(loss.item() * accumulation_steps)
+            loss = combined_loss(output, target)
+
+            loss.backward()
+            
+            # Gradient clipping for training stability
+            torch.nn.utils.clip_grad_norm_(network.parameters(), max_norm=1.0)
+
+            optimizer.step()
+            scheduler.step(i + len(loss_list) / len(dataloader_train))
+
+            loss_list.append(loss.item())
 
             # writing the stats to wandb
             if use_wandb and (i+1) % print_stats_at == 0:
@@ -188,9 +165,7 @@ def train(seed, testset_ratio, validset_ratio, data_path, results_path, early_st
             # plotting
             if (i + 1) % plot_at == 0:
                 print(f"Plotting images, current update {i + 1}")
-                # Convert to float32 for matplotlib compatibility (mixed precision may produce float16)
+                plot(input.cpu().numpy(), target.detach().cpu().numpy(), output.detach().cpu().numpy(), plotpath, i)
-                plot(input.float().cpu().numpy(), target.detach().float().cpu().numpy(), 
-                     output.detach().float().cpu().numpy(), plotpath, i)
 
             # evaluating model every validate_at sample
             if (i + 1) % validate_at == 0: