Add the config about using checkpoint in PVTv2.

README.md

@@ -177,3 +177,7 @@ python main_for_image.py --config configs/icod_train.py --pretrained --model-nam
 
 1. Pretrain on COD10K-TR: `python main_for_image.py --config configs/icod_pretrain.py --info pretrain --model-name PvtV2B5_ZoomNeXt --pretrained`
 2. Finetune on MoCA-Mask-TR: `python main_for_video.py --config configs/vcod_finetune.py --info finetune --model-name videoPvtV2B5_ZoomNeXt --load-from <PRETAINED_WEIGHT>`
+
+> [!note]
+> If you meets the OOM problem, you can try to reduce the batch size or switch on the `--use-checkpoint` flag:
+> `python main_for_image.py/main_for_video.py <your config> --use-checkpoint`

main_for_image.py

@@ -345,6 +345,7 @@ def parse_cfg():
     )
     parser.add_argument("--evaluate", action="store_true")
     parser.add_argument("--save-results", action="store_true")
+    parser.add_argument("--use-checkpoint", action="store_true")
     parser.add_argument("--info", type=str)
     args = parser.parse_args()
 
@@ -382,7 +383,7 @@ def main():
     model_class = model_zoo.__dict__.get(cfg.model_name)
     assert model_class is not None, "Please check your --model-name"
     model_code = inspect.getsource(model_class)
-    model = model_class(num_frames=1, pretrained=cfg.pretrained)
+    model = model_class(num_frames=1, pretrained=cfg.pretrained, use_checkpoint=cfg.use_checkpoint)
     LOGGER.info(model_code)
     model.to(cfg.device)
 

main_for_video.py

@@ -467,6 +467,7 @@ def parse_cfg():
     )
     parser.add_argument("--evaluate", action="store_true")
     parser.add_argument("--save-results", action="store_true")
+    parser.add_argument("--use-checkpoint", action="store_true")
     parser.add_argument("--info", type=str)
     args = parser.parse_args()
 
@@ -504,7 +505,7 @@ def main():
     model_class = model_zoo.__dict__.get(cfg.model_name)
     assert model_class is not None, "Please check your --model-name"
     model_code = inspect.getsource(model_class)
-    model = model_class(num_frames=cfg.num_frames, pretrained=cfg.pretrained)
+    model = model_class(num_frames=cfg.num_frames, pretrained=cfg.pretrained, use_checkpoint=cfg.use_checkpoint)
     LOGGER.info(model_code)
     model.to(cfg.device)
 

methods/zoomnext/zoomnext.py

@@ -86,7 +86,9 @@ def get_grouped_params(self):
 
 
 class RN50_ZoomNeXt(_ZoomNeXt_Base):
-    def __init__(self, pretrained=True, num_frames=1, input_norm=True, mid_dim=64, siu_groups=4, hmu_groups=6):
+    def __init__(
+        self, pretrained=True, num_frames=1, input_norm=True, mid_dim=64, siu_groups=4, hmu_groups=6, **kwargs
+    ):
         super().__init__()
         self.encoder = timm.create_model(
             model_name="resnet50", features_only=True, out_indices=range(5), pretrained=False
@@ -297,7 +299,7 @@ def get_grouped_params(self):
 
 
 class EffB1_ZoomNeXt(_ZoomNeXt_Base):
-    def __init__(self, pretrained, num_frames=1, input_norm=True, mid_dim=64, siu_groups=4, hmu_groups=6):
+    def __init__(self, pretrained, num_frames=1, input_norm=True, mid_dim=64, siu_groups=4, hmu_groups=6, **kwargs):
         super().__init__()
         self.set_backbone(pretrained)