I tested your pre-trained SOLOV2... model that everything is perfect.
But when I train your code with the default settings of solov2_r101_dcn_fpn_8gpu_3x.py with one GPU (we do not have 8 GPUs), the loss gets stuck at some point.
To help me solve the above problems, may I ask you for the log files during the training process if it is convenient?
Dear contributor,
I tested your pre-trained SOLOV2... model that everything is perfect.
But when I train your code with the default settings of solov2_r101_dcn_fpn_8gpu_3x.py with one GPU (we do not have 8 GPUs), the loss gets stuck at some point.
To help me solve the above problems, may I ask you for the log files during the training process if it is convenient?
Kind Regards