src/aiunn/finetune.py

@@ -12,7 +12,7 @@ import numpy as np
 from torch import nn
 from torch.utils.data import random_split, DataLoader
 from aiia import AIIA, AIIAConfig, AIIABase, AIIABaseShared, AIIAmoe, AIIAchunked, AIIArecursive
-from torch.cuda.amp import autocast, GradScaler
+from torch.amp import autocast, GradScaler
 from tqdm import tqdm
 from torch.utils.checkpoint import checkpoint
 
@@ -87,7 +87,6 @@ def finetune_model(model: AIIA, datasets: list[str], batch_size=1, epochs=10, ac
     )
     
     device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
-    # Fix: Pass the current device index (an integer) rather than a torch.device without index.
     if device.type == 'cuda':
         current_device = torch.cuda.current_device()
         torch.cuda.set_per_process_memory_fraction(0.95, device=current_device)
@@ -109,7 +108,6 @@ def finetune_model(model: AIIA, datasets: list[str], batch_size=1, epochs=10, ac
             high_res = batch['high_res'].to(device)
             with autocast():
                 if use_checkpoint:
-                    # Use checkpointing to save intermediate activations if needed.
                     outputs = checkpoint(lambda x: model(x), low_res)
                 else:
                     outputs = model(low_res)
@@ -120,7 +118,6 @@ def finetune_model(model: AIIA, datasets: list[str], batch_size=1, epochs=10, ac
                 scaler.step(optimizer)
                 scaler.update()
                 optimizer.zero_grad()
-        # Handle leftover gradients
         if (i % accumulation_steps) != 0:
             scaler.step(optimizer)
             scaler.update()
@@ -149,14 +146,24 @@ def finetune_model(model: AIIA, datasets: list[str], batch_size=1, epochs=10, ac
     return model
 
 def main():
-    BATCH_SIZE = 2           # Use a batch size of 2.
+    BATCH_SIZE = 2
-    ACCUMULATION_STEPS = 8   # Accumulate gradients to simulate a larger batch.
+    ACCUMULATION_STEPS = 8
-    USE_CHECKPOINT = True   # Set to True to enable gradient checkpointing if needed.
+    USE_CHECKPOINT = True
-    
+
     model = AIIABase.load("/root/vision/AIIA/AIIA-base-512")
+    
+    # Upsample output if a 'chunked_' attribute exists, ensuring spatial dimensions match the high resolution images.
     if hasattr(model, 'chunked_'):
         model.add_module('final_upsample', nn.Upsample(scale_factor=2, mode='bilinear'))
     
+    # Append a final convolutional layer using values from model.config.
+    # This converts the hidden feature maps (512 channels) to the 3 required output channels.
+    final_conv = nn.Conv2d(model.config.hidden_size, model.config.num_channels, kernel_size=3, padding=1)
+    model.add_module('final_layer', final_conv)
+    
+    print("Modified model architecture:")
+    print(model.config)
+    
     finetune_model(
         model=model,
         datasets=[