There are no logs outputs information like loss etc. during the training Process

I use

python tools/train.py -f exps/example/yolox_voc/yolox_voc_s.py -d 8 -b 128 --fp16 -o -c weights/yolox_s.pth

for multi GPUS training but there are no logs outputting

`
2022-07-06 14:44:58 | INFO | yolox.core.trainer:203 - ---> start train epoch1

2022-07-06 14:45:07 | INFO | yolox.core.trainer:356 - Save weights to ./YOLOX_outputs/yolox_voc_s

2022-07-06 14:45:07 | INFO | yolox.core.trainer:203 - ---> start train epoch2

2022-07-06 14:45:14 | INFO | yolox.core.trainer:356 - Save weights to ./YOLOX_outputs/yolox_voc_s

`

while is all fine using only 1 GPU

python tools/train.py -f exps/example/yolox_voc/yolox_voc_s.py -d 0 -b 4 --fp16 -o -c weights/yolox_s.pth

like:

`
2022-07-06 15:07:47 | INFO | yolox.core.trainer:203 - ---> start train epoch1

2022-07-06 15:07:50 | INFO | yolox.core.trainer:253 - epoch: 1/300, iter: 10/128, mem: 10091Mb, iter_time: 0.328s, data_time: 0.001s, total_loss: 15.5, iou_loss: 3.6, l1_loss: 0.0, conf_loss: 9.6, cls_loss: 2.3, lr: 3.815e-06, size: 640, ETA: 3:29:45

2022-07-06 15:07:53 | INFO | yolox.core.trainer:253 - epoch: 1/300, iter: 20/128, mem: 10091Mb, iter_time: 0.311s, data_time: 0.002s, total_loss: 15.7, iou_loss: 3.5, l1_loss: 0.0, conf_loss: 10.1, cls_loss: 2.1, lr: 1.526e-05, size: 768, ETA: 3:24:10
`

Reproduce the commands in tools/train.py for single-GPU and multi-GPU training, then compare the logging path in yolox.core.trainer, especially the lines shown in the messages. Confirm that multi-GPU training reports loss and iteration details like the single-GPU run; the issue is done when those training metrics appear during multi-GPU execution.