diff --git a/gan.sh b/gan.sh
index 4767ca5..5065d5c 100644
--- a/gan.sh
+++ b/gan.sh
@@ -2,7 +2,7 @@
 python3 -m minamo.train --epochs 10 --resume true
 python3 -m minamo.train --epochs 10 --resume true --train "datasets/minamo-dataset-1.json" --validate "datasets/minamo-eval-1.json"
 python3 -m minamo.train --epochs 10 --resume true
-python3 -m ginka.train --epochs 70 --resume true
+python3 -m ginka.train --epochs 30 --resume true
 python3 -m ginka.validate
 # 训练完毕，处理数据
 mv "minamo-dataset.json" "datasets/minamo-dataset-$1.json"
diff --git a/ginka/train.py b/ginka/train.py
index c339629..9b0dbf7 100644
--- a/ginka/train.py
+++ b/ginka/train.py
@@ -75,7 +75,7 @@ def train():
             feat_vec = torch.cat([target_vision_feat, target_topo_feat], dim=-1).to(device).squeeze(1)
             # 前向传播
             optimizer.zero_grad()
-            noise = torch.randn((BATCH_SIZE, 1, 32, 32))
+            noise = torch.randn((target.shape[0], 1, 32, 32)).to(device)
             _, output_softmax = model(noise, feat_vec)
             
             # 计算损失
@@ -108,7 +108,8 @@ def train():
                     feat_vec = torch.cat([target_vision_feat, target_topo_feat], dim=-1).to(device).squeeze(1)
                     
                     # 前向传播
-                    output, output_softmax = model(feat_vec)
+                    noise = torch.randn((target.shape[0], 1, 32, 32)).to(device)
+                    output, output_softmax = model(noise, feat_vec)
                     print(torch.argmax(output, dim=1)[0])
                     
                     # 计算损失
diff --git a/ginka/validate.py b/ginka/validate.py
index 89ef941..750707f 100644
--- a/ginka/validate.py
+++ b/ginka/validate.py
@@ -108,7 +108,8 @@ def validate():
             target_topo_feat = batch["target_topo_feat"].to(device)
             feat_vec = torch.cat([target_vision_feat, target_topo_feat], dim=-1).to(device).squeeze(1)
             # 前向传播
-            output, output_softmax = model(feat_vec)
+            noise = torch.randn((target.shape[0], 1, 32, 32)).to(device)
+            output, output_softmax = model(noise, feat_vec)
             map_matrix = torch.argmax(output, dim=1)
             
             for matrix in map_matrix[:].cpu():
diff --git a/minamo/dataset.py b/minamo/dataset.py
index 079c4d4..fb99639 100644
--- a/minamo/dataset.py
+++ b/minamo/dataset.py
@@ -29,9 +29,9 @@ class MinamoDataset(Dataset):
         map1_probs = F.one_hot(torch.LongTensor(item['map1']), num_classes=32).permute(2, 0, 1).float()  # [32, H, W]
         map2_probs = F.one_hot(torch.LongTensor(item['map2']), num_classes=32).permute(2, 0, 1).float()  # [32, H, W]
         
-        min_main = random.uniform(0.7, 1)
-        max_main = random.uniform(0.9, 1)
-        epsilon = random.uniform(0, 0.3)
+        min_main = random.uniform(0.6, 1)
+        max_main = random.uniform(0.8, 1)
+        epsilon = random.uniform(0, 0.4)
         
         map1_probs = random_smooth_onehot(map1_probs, min_main, max_main, epsilon)
         map2_probs = random_smooth_onehot(map2_probs, min_main, max_main, epsilon)