Bug: YOLOX-PAI training evaluation giving 0 mAPs

[junxiant]

Describe the bug
YOLOX-PAI object detection training.
The evaluation during training process gives 0 for all mAPs.

easycv - INFO - Epoch(val) [10][7393] CocoDetectionEvaluator_DetectionBoxes_Precision/mAP: 0.0000, DetectionBoxes_Precision/mAP: 0.0000, DetectionBoxes_Precision/mAP@.50IOU: 0.0000, DetectionBoxes_Precision/mAP@.75IOU: 0.0000, DetectionBoxes_Precision/mAP (small): 0.0000, DetectionBoxes_Precision/mAP (medium): 0.0000, DetectionBoxes_Precision/mAP (large): 0.0000, DetectionBoxes_Recall/AR@1: 0.0000, DetectionBoxes_Recall/AR@10: 0.0000, DetectionBoxes_Recall/AR@100: 0.0000, DetectionBoxes_Recall/AR@100 (small): 0.0000, DetectionBoxes_Recall/AR@100 (medium): 0.0000, DetectionBoxes_Recall/AR@100 (large): 0.0000

To Reproduce

• What command or script did you run?

bash tools/dist_train.sh
${NUM_GPUS}
${CONFIG_PATH}
--work_dir ${WORK_DIR}

• Did you make any modifications on the code or config? Did you understand what you have modified?
  No modifications was made to the code/config. Strictly followed the yolox.md tutorial.

• What dataset did you use?
  COCO dataset, following the readme.

Environment
Followed the instructions to create an ev environment and pip installing requirements.txt