Update test_config.py

Zplusdragon · May 31, 2023 · 735200f · 735200f
1 parent da92d93
commit 735200f
Showing 1 changed file with 20 additions and 29 deletions.
diff --git a/test_config.py b/test_config.py
@@ -13,31 +13,25 @@
 
 def Test_parse_args():
     parser = argparse.ArgumentParser()
-
-    parser.add_argument('--model_path', type=str, default='/home/dancer/Papercode/PLIPv2/log_AA/MA__VAP')
-    parser.add_argument('--image_path', type=str, default='/home/dancer/Papercode/FISNet_deit_bert/data/cuhk')
+    parser.add_argument('--model_path', type=str, default='checkpoints/PLIP_MRN50.pth.tar')
+    parser.add_argument('--image_path', type=str, default='data/cuhk')
     parser.add_argument('--test_path', type=str,
-                        default='/home/dancer/Papercode/FISNet_deit_bert/data/cuhk_annotations/CUHK-PEDES-test.json',
+                        default='data/CUHK-PEDES-test.json',
                         help='path for test annotation json file')
 # ***********************************************************************************************************************
 # 设置模型backbone的类型和参数
     parser.add_argument('--plip_model', type=str, default='MResNet_BERT')
-    parser.add_argument('--img_backbone', type=str, default='Swin_Base_BERT',
+    parser.add_argument('--img_backbone', type=str, default='ModifiedResNet',
                         help="ResNet:xxx, ModifiedResNet, ViT:xxx")
     parser.add_argument('--txt_backbone', type=str, default="bert-base-uncased")
     parser.add_argument('--img_dim', type=int, default=768, help='dimension of image embedding vectors')
     parser.add_argument('--text_dim', type=int, default=768, help='dimension of text embedding vectors')
-    parser.add_argument('--patch_size', type=int, default=16, help='Just for ViT model')
     parser.add_argument('--layers', type=list, default=[3, 4, 6, 3], help='Just for ModifiedResNet model')
     parser.add_argument('--heads', type=int, default=8, help='Just for ModifiedResNet model')
-
-    parser.add_argument('--content_mask_ratio', type=float, default=0.5, help='dimension of lstm hidden states')
-    parser.add_argument('--others_mask_ratio', type=float, default=0.1, help='number of layers in lstm')
     parser.add_argument('--height', type=int, default=256)
     parser.add_argument('--width', type=int, default=128)
 
     # 设置超参数
-    parser.add_argument('--num_epoches', type=int, default=70)
     parser.add_argument('--batch_size', type=int, default=128)
     parser.add_argument('--num_workers', type=int, default=4)
     parser.add_argument('--device',type=str,default="cuda:0")
@@ -91,31 +85,28 @@ def Test_main(args):
     ac_t2i_top10_best = 0.0
     mAP_best = 0.0
     best = 0
-    #dst_best = args.best_dir + "/model_best" + ".pth"
     model = Create_PLIP_Model(args).to(device)
 
-    for i in range(29, 41):
-        model_file = os.path.join(args.model_path, str(i+1))+".pth.tar"
-        print(model_file)
-        if os.path.isdir(model_file):
-            continue
-        checkpoint = torch.load(model_file,map_location='cpu')
-        model.image_encoder.load_state_dict(checkpoint["ImgEncoder_state_dict"],strict=False)
-        model.text_encoder.load_state_dict(checkpoint["TxtEncoder_state_dict"])
-        ac_top1_t2i, ac_top5_t2i, ac_top10_t2i, mAP = test(image_test_loader,text_test_loader, model)
-        if ac_top1_t2i > ac_t2i_top1_best:
-            ac_t2i_top1_best = ac_top1_t2i
-            ac_t2i_top5_best = ac_top5_t2i
-            ac_t2i_top10_best = ac_top10_t2i
-            mAP_best = mAP
-            best = i
-            #shutil.copyfile(model_file, dst_best)
-
+    model_file = args.model_path
+    print(model_file)
+    if os.path.isdir(model_file):
+        continue
+    checkpoint = torch.load(model_file,map_location='cpu')
+    model.image_encoder.load_state_dict(checkpoint["ImgEncoder_state_dict"])
+    model.text_encoder.load_state_dict(checkpoint["TxtEncoder_state_dict"])
+    ac_top1_t2i, ac_top5_t2i, ac_top10_t2i, mAP = test(image_test_loader,text_test_loader, model)
+    if ac_top1_t2i > ac_t2i_top1_best:
+        ac_t2i_top1_best = ac_top1_t2i
+        ac_t2i_top5_best = ac_top5_t2i
+        ac_t2i_top10_best = ac_top10_t2i
+        mAP_best = mAP
+        best = i
+
     print('Epo{}:  {:.5f}  {:.5f}  {:.5f}  {:.5f}'.format(
             best+1, ac_t2i_top1_best, ac_t2i_top5_best, ac_t2i_top10_best, mAP_best))
 
 import warnings
 warnings.filterwarnings("ignore")
 if __name__ == '__main__':
     args = Test_parse_args()
-    Test_main(args)
+    Test_main(args)