Evaluation Metrics Show Zero Values Despite Correct Label Loading in Training

Hello, may I ask why when I am training, it keeps showing：
Epoch gpu_mem box obj cls pixl area dist total labels img_size
0/49 0.912G 0.1011 0.1259 0.0476 0.005673 0 0 0.2758 171 1024: 100%|██| 809/809 [02:12<00:00, 6.12it/s]
Class Images Labels P R mAP@.5 mAP@.5:.95 BPR Occupy: 100%|██████| 69/69 [00:02<00:00, 24.53it/s]
all 548 0 0 0 0 0 nan 1
I've checked that my val labels are reading correctly, but why does the assessment keep showing 0 for labels and 0 for other metrics?

The issue names no files, command, dataset format, or tests. Start by collecting the training and validation configuration and reproducing the reported epoch output; compare the loaded validation-label count with the metric inputs. Done means the cause is documented and the validation summary reports the expected labels and metrics.

Written by the indexing model from the issue text.