when I train my own dataset.P,R,map is zero

[futureflsl]

I annotated some images,and find that this project train can successfully start,but P,R,map is zero all the time,I train 200 epoch.

Epoch gpu_mem box obj cls total labels img_size
64/199 3.17G 0.03754 0.01328 0 0.05082 11 640: 100%|█| 21/21 [00:02<00:00, 9.
Class Images Labels P R [email protected] [email protected]:.95: 100%|█| 2/2 [00:00<00:00,
all 19 0 0 0 0 0

 Epoch   gpu_mem       box       obj       cls     total    labels  img_size
65/199     3.17G   0.03064    0.0133         0   0.04394         9       640: 100%|█| 21/21 [00:02<00:00,  9.
           Class     Images     Labels          P          R     [email protected] [email protected]:.95: 100%|█| 2/2 [00:00<00:00,
             all         19          0          0          0          0          0

 Epoch   gpu_mem       box       obj       cls     total    labels  img_size
66/199     3.17G   0.03182   0.01224         0   0.04405        12       640:  62%|▌| 13/21 [00:01<00:00,  9.    66/199     3.17G   0.03182   0.01224         0   0.04405        12       640:  62%|▌| 13/21 [00:01<00:00,  9.

[XinzeLee]

From your training history, I can tell that you only adopt around 200 images for training. This is not a domain transfer or few shot repo. Please add more data for training.

[lemon1022]

@futureflsl Did you solve it?