WARNING [d2.data.common]: Failed to apply `_map_func` for idx: N

[JihoChoi]

When I run the evaluation script, I get WARNINGs that say Failed to apply _map_func for idx: N. Do anyone now why this happens? I am wondering if the warning is something that needs to be fixed or not.

Thank you very much.

Run

python train_net.py \
    --num-gpus 1 \
    --config-file configs/zero_shot/clipseg_voc.yaml \
    --eval-only MODEL.WEIGHTS ./weights/clipseg_ft_VA_L_F_voc.pth

Warning messages

RNING [02/13 00:12:23 d2.data.common]: Failed to apply `_map_func` for idx: 905, retry count: 8
WARNING [02/13 00:12:25 d2.data.common]: Failed to apply `_map_func` for idx: 916, retry count: 3
[02/13 00:12:27 d2.evaluation.evaluator]: Inference done 911/1209. Dataloading: 0.0064 s/iter. Inference: 0.1654 s/iter. Eval: 0.0099 s/iter. Total: 0.1817 s/iter. ETA=0:00:54
WARNING [02/13 00:12:32 d2.data.common]: Failed to apply `_map_func` for idx: 951, retry count: 3
WARNING [02/13 00:12:32 d2.data.common]: Failed to apply `_map_func` for idx: 951, retry count: 4
WARNING [02/13 00:12:32 d2.data.common]: Failed to apply `_map_func` for idx: 951, retry count: 5
[02/13 00:12:32 d2.evaluation.evaluator]: Inference done 937/1209. Dataloading: 0.0064 s/iter. Inference: 0.1657 s/iter. Eval: 0.0099 s/iter. Total: 0.1821 s/iter. ETA=0:00:49
[02/13 00:12:38 d2.evaluation.evaluator]: Inference done 964/1209. Dataloading: 0.0064 s/iter. Inference: 0.1660 s/iter. Eval: 0.0099 s/iter. Total: 0.1823 s/iter. ETA=0:00:44
WARNING [02/13 00:12:40 d2.data.common]: Failed to apply `_map_func` for idx: 991, retry count: 3
WARNING [02/13 00:12:40 d2.data.common]: Failed to apply `_map_func` for idx: 991, retry count: 4
[02/13 00:12:43 d2.evaluation.evaluator]: Inference done 990/1209. Dataloading: 0.0064 s/iter. Inference: 0.1664 s/iter. Eval: 0.0099 s/iter. Total: 0.1828 s/iter. ETA=0:00:40
[02/13 00:12:48 d2.evaluation.evaluator]: Inference done 1016/1209. Dataloading: 0.0064 s/iter. Inference: 0.1667 s/iter. Eval: 0.0099 s/iter. Total: 0.1831 s/iter. ETA=0:00:35
WARNING [02/13 00:12:52 d2.data.common]: Failed to apply `_map_func` for idx: 1052, retry count: 3
[02/13 00:12:53 d2.evaluation.evaluator]: Inference done 1043/1209. Dataloading: 0.0064 s/iter. Inference: 0.1669 s/iter. Eval: 0.0099 s/iter. Total: 0.1833 s/iter. ETA=0:00:30
[02/13 00:12:58 d2.evaluation.evaluator]: Inference done 1070/1209. Dataloading: 0.0064 s/iter. Inference: 0.1670 s/iter. Eval: 0.0099 s/iter. Total: 0.1834 s/iter. ETA=0:00:25
WARNING [02/13 00:13:00 d2.data.common]: Failed to apply `_map_func` for idx: 1094, retry count: 3
[02/13 00:13:03 d2.evaluation.evaluator]: Inference done 1097/1209. Dataloading: 0.0064 s/iter. Inference: 0.1672 s/iter. Eval: 0.0099 s/iter. Total: 0.1836 s/iter. ETA=0:00:20
WARNING [02/13 00:13:08 d2.data.common]: Failed to apply `_map_func` for idx: 1139, retry count: 3
[02/13 00:13:08 d2.evaluation.evaluator]: Inference done 1124/1209. Dataloading: 0.0064 s/iter. Inference: 0.1673 s/iter. Eval: 0.0099 s/iter. Total: 0.1838 s/iter. ETA=0:00:15
[02/13 00:13:13 d2.evaluation.evaluator]: Inference done 1151/1209. Dataloading: 0.0064 s/iter. Inference: 0.1674 s/iter. Eval: 0.0099 s/iter. Total: 0.1838 s/iter. ETA=0:00:10

[kellyiss]

This warning is from the dataloader of Detectron2. It occurs because some objects in the dataset do not have part masks. We address this by removing redundant predictions after model inference. Additionally, filtering these test annotations before model inference can help mitigate the issue as well.