Kui Xu

Error in CuDNN: CUDNN_STATUS_BAD_PARAM (cudnnSetConvolutionNdDescriptor)

4

I encounter these error messages, when run cifar10 with 2 GPUs. Command line is `CUDA_VISIBLE_DEVICES=0,1 th main.lua -depth 50 -batchSize 1024 -nGPU 2 -nThreads 8 -shareGradInput true -dataset cifar10` ....

ImportError: No module named dataset_loaders.images.camvid

14

Looking forward to the dataset_loaders. ``` Traceback (most recent call last): File "train.py", line 16, in from data_loader import load_data File "/data/kuixu/exper/fcn/FC-DenseNet/data_loader.py", line 1, in from dataset_loaders.images.camvid import CamvidDataset ImportError:...

demo_convnet.py run with error for the reason of libsmat_cudnn.so

1

`demo_convnet.py` run with error for the reason of `libsmat_cudnn.so` code/libs/smat/py/demo_convnet.py $ demo_convnet.py --device 1 code/libs/smat/build/release/bin/libsmat_cudnn.so: cannot open shared object file: No such file or directory and then I found that...

How to debug the error of 'vmtouch: FATAL: daemon shut down unexpectedly'? The command is `vmtouch -f -w -t -l -d -m 500G /path/to/a/448G/file`.