smh

when I use this docker image, an error happened: ``` E0407 17:29:21.355960 120 common.cpp:114] Cannot create Cublas handle. Cublas won't be available. E0407 17:29:21.356129 120 common.cpp:121] Cannot create Curand generator....

> when I use this docker image, an error happened: > > ``` > E0407 17:29:21.355960 120 common.cpp:114] Cannot create Cublas handle. Cublas won't be available. > E0407 17:29:21.356129 120...

same error. may be you could have a look at this : [issues72](https://github.com/ChenRocks/UNITER/issues/72#issuecomment-814809422)

I have a question, the features extracted from resnet is not same as the downloaded `.h5` data, so, would the data distribution mismatch between two feature extractor network cause not...

You should select the InputAsShotZeroed as the input data.

ok, thanks

My server contains 8 P40 GPUs. I tried just using one GPU (`cuda:0`) and same error happened. I used ``` python os.environ['CUDA_VISIBLE_DEVICE']="0" # OR torch.cuda.set_device(0) ``` to use `cuda:0` only....

Sorry, not yet ... It may be caused by the pytorch version, gcc version, but I am not sure. By the way, my pytorch version is 1.0.2