fix: maskGIT

ginka/train_maskGIT.py

@@ -93,9 +93,9 @@ def train():
 
     # 用于生成图片
     tile_dict = dict()
-    for file in os.listdir('tiles'):
+    for file in os.listdir('tiles2'):
         name = os.path.splitext(file)[0]
-        tile_dict[name] = cv2.imread(f"tiles/{file}", cv2.IMREAD_UNCHANGED)
+        tile_dict[name] = cv2.imread(f"tiles2/{file}", cv2.IMREAD_UNCHANGED)
         
     if args.resume:
         data_ginka = torch.load(args.state_ginka, map_location=device)
@@ -111,30 +111,32 @@ def train():
     for epoch in tqdm(range(args.epochs), desc="VAE Training", disable=disable_tqdm):
         loss_total = torch.Tensor([0]).to(device)
         
-        # for batch in tqdm(dataloader, leave=False, desc="Epoch Progress", disable=disable_tqdm):
+        for batch in tqdm(dataloader, leave=False, desc="Epoch Progress", disable=disable_tqdm):
-        #     target_map = batch["target_map"].to(device)
+            target_map = batch["target_map"].to(device)
-        #     cond = batch["val_cond"].to(device)
+            cond = batch["val_cond"].to(device)
-        #     B, H, W = target_map.shape
+            B, H, W = target_map.shape
-        #     target_map = target_map.view(B, H * W)
+            target_map = target_map.view(B, H * W)
             
-        #     # 1. 随机采样掩码比例 r (遵循余弦调度效果更好)
+            # 1. 随机采样掩码比例 r (遵循余弦调度效果更好)
-        #     r = torch.rand(B).to(device)
+            r = torch.rand(B).to(device)
-        #     r = torch.cos(r * math.pi / 2).unsqueeze(1) # 产生更多高掩码比例的样本
+            r = torch.cos(r * math.pi / 2).unsqueeze(1) # 产生更多高掩码比例的样本
             
-        #     # 2. 生成掩码矩阵
+            # 2. 生成掩码矩阵
-        #     masks = torch.rand(target_map.shape).to(device) < r
+            masks = torch.rand(target_map.shape).to(device) < r
-        #     masked_input = target_map.clone()
+            masked_input = target_map.clone()
-        #     masked_input[masks] = MASK_TOKEN # 填充为 [MASK] 标记
+            masked_input[masks] = MASK_TOKEN # 填充为 [MASK] 标记
             
-        #     logits = model(masked_input, cond)
+            logits = model(masked_input, cond)
             
-        #     loss = F.cross_entropy(logits.permute(0, 2, 1), target_map, reduction='none')
+            loss = F.cross_entropy(logits.permute(0, 2, 1), target_map, reduction='none')
-        #     loss = (loss * masks).sum() / (masks.sum() + 1e-6)
+            loss = (loss * masks).sum() / (masks.sum() + 1e-6)
             
-        #     optimizer.zero_grad()
+            optimizer.zero_grad()
-        #     loss.backward()
+            loss.backward()
-        #     optimizer.step()
+            optimizer.step()
-        #     loss_total += loss.detach()
+            loss_total += loss.detach()
+            
+        scheduler.step()
                 
         avg_loss = loss_total.item() / len(dataloader)
         tqdm.write(

ginka/train_transformer.py

@@ -102,9 +102,9 @@ def train():
 
     # 用于生成图片
     tile_dict = dict()
-    for file in os.listdir('tiles'):
+    for file in os.listdir('tiles2'):
         name = os.path.splitext(file)[0]
-        tile_dict[name] = cv2.imread(f"tiles/{file}", cv2.IMREAD_UNCHANGED)
+        tile_dict[name] = cv2.imread(f"tiles2/{file}", cv2.IMREAD_UNCHANGED)
         
     if args.resume:
         data_ginka = torch.load(args.state_ginka, map_location=device)
@@ -196,6 +196,7 @@ def train():
                 color = (255, 255, 255)  # 白色
                 vline = np.full((416, gap, 3), color, dtype=np.uint8)  # 垂直分割线
                 # 地图重建展示
+                vae.teacher_forcing()
                 for batch in tqdm(dataloader_val, desc="Validating generator.", leave=False, disable=disable_tqdm):
                     target_map = batch["target_map"].to(device)
                     B, H, W = target_map.shape
@@ -218,10 +219,10 @@ def train():
                     idx += 1
                 
                 # 随机采样
+                vae.autoregressive()
                 for i in range(0, 8):
                     z = torch.randn(1, LATENT_DIM).to(device)
                     
-                    vae.autoregressive()
                     fake_logits = vae.decoder(z, torch.zeros(1, 169).to(device))
                     fake_map = fake_logits[:,0:169].view(-1, 13, 13).cpu().numpy()
                     fake_img = matrix_to_image_cv(fake_map[0], tile_dict)