BiSeNet
BiSeNet copied to clipboard
use for kitti, how to train or finetune,thanks!
Dear author,
If I want to use the model to do the training in the Kitti dataset. So we have the different category number, for the sityscapes ,it is 20 . For Kitti, for examplt, the category number is 6. So how to o the training work? add a conv2d(20, 6)? or anything else.
I use the command below:L torchrun --nproc_per_node=2 train_amp.py --finetune-from pretrain_weight/model_final_v2_city.pth
But it shows that missing a lot of key : I just set the category num to 5.
Maybe you are using unmatched config file and finetune weight file.
Close this, since this is not activate for long time, and likely the problem is solved.