perf(bbox_regression): 使用Adam；调整targets精度为float32

Author: zjZSTU

py/bbox_regression.py

@@ -49,8 +49,9 @@ def train_model(data_loader, feature_model, model, criterion, optimizer, lr_sche
 
         # Iterate over data.
         for inputs, targets in data_loader:
+            print(targets.dtype)
             inputs = inputs.to(device)
-            targets = targets.to(device)
+            targets = targets.float().to(device)
 
             features = feature_model.features(inputs)
             features = torch.flatten(features, 1)
@@ -113,7 +114,8 @@ def get_model(device=None):
     model.to(device)
 
     criterion = nn.MSELoss()
-    optimizer = optim.SGD(model.parameters(), lr=1e-3, momentum=0.9, weight_decay=1e-3)
+    # optimizer = optim.SGD(model.parameters(), lr=1e-3, momentum=0.9, weight_decay=1e-3)
+    optimizer = optim.Adam(model.parameters(), lr=1e-3, weight_decay=1e-3)
     lr_scheduler = optim.lr_scheduler.StepLR(optimizer, step_size=7, gamma=0.1)
 
     loss_list = train_model(data_loader, feature_model, model, criterion, optimizer, lr_scheduler, device=device,