Unable to train custom dataset in GPU

Community Help
1

yolo detect train model=yolov8n.pt epochs=100 imgsz=640

in Gpu
Starting training for 1 epochs…

  Epoch    GPU_mem   box_loss   cls_loss   dfl_loss  Instances       Size
    1/1      3.48G        nan        nan        nan          1        640: 100%|██████████| 22/22 [01:49<00:00,  4.96s/it]
             Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 11/11 [00:38<00:00,  3.49s/it]    
               all        169        181          0          0          0          0

in Cpu
Starting training for 1 epochs…

  Epoch    GPU_mem   box_loss   cls_loss   dfl_loss  Instances       Size
    1/1         0G      1.163      2.564      1.624          4        640: 100%|██████████| 22/22 [05:49<00:00, 15.89s/it]
             Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 11/11 [02:06<00:00, 11.48s/it]
               all        169        181      0.204      0.654       0.23     0.0869
2

Hi @Jaishankar_Vetrivel

Can you explain and add more detail about what you are having issues with?

3

It worked.

yolo task=detect mode=train epochs=50 data=data.yaml model=yolov8m.pt imgsz=640 amp=False batch=8 patience=50

We specified amp=False to resolve the issue.
Thanks

4

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.