Train your own model with ssd_mobilenet_v1_coco

[8787878877]

Hello, I would like to ask, I used the detection.ipynb program to run on jetson nano.
This model uses ssd_mobilenet_v1_coco for migration learning. Before this, I have tested that the sample program can indeed run on nano, but when I run it with the model I have trained, the accuracy and frame cannot be displayed. Is it related to my training data or is there something else? the reason.
Thank you