lt573152873

resnext-101-64f-kinetics-hmdb51_split1.pth model question?

2

when I use resnext-101-64f-kinetics-hmdb51_split1.pth model, dataset is hmdb51. should I choose which parameter for opt ？

during training, Occu_mask whether there is gradient return? when occu_mask is zeros, the loss is small; if need occu_mask -> occu_mask.detach() ?

I use LS3D-W dataset , when I training, the loss is lower .. expamle: total_loss : 1.1131469011306763 classification: 0.70500928 bbox: 0.07543718814849854 landmark: 0.6654008626937866 . Check whether the fault is rectified.

actual images result bad why?

1

For the actual images, The results of RTM3D are better than those of RTS3D. why?

Can I deploy mx-maskrcnn in the ubuntu14.04？

1

i use cuda-7.0. when i make， fatal error :cuda_fp16.h: No such file or directory. when i python train_mnist.py --gpus 0 --network lenet , OSError: libcudart.so.8.0: cannot open shared object file...

bash /scripts/train_alternate.sh after a long time , the computer is automatically restarted; why?

why RPNLogLoss=nan and RPN1Loss=nan?

3

I use one GPU; after bash /scripts/train_alternate.sh ; Just started showing the correct information； when Batch[680], showing RPNLogLoss=nan and RPN1Loss=nan; why?