Open hainguyenct opened 7 years ago
Hello, I have data with Tensor and a model for example, data:
local x = torch.Tensor(opt.batchSize,trainData.data:size(2), trainData.data:size(3), trainData.data:size(4)) -- local yt = torch.Tensor(opt.batchSize)
model: _CNN:add(nn.SpatialConvolution(channelimage, nstates,filtsize, filtsize)) I converted this data into cuda: x = x:cuda() yt = yt:cuda()
CNN=CNN:cuda()
However, the speed seems to be unchanged when using cuda and without cuda. I do not know why?
Hello, I have data with Tensor and a model for example, data:
local x = torch.Tensor(opt.batchSize,trainData.data:size(2), trainData.data:size(3), trainData.data:size(4)) -- local yt = torch.Tensor(opt.batchSize)
model: _CNN:add(nn.SpatialConvolution(channelimage, nstates,filtsize, filtsize)) I converted this data into cuda: x = x:cuda() yt = yt:cuda()
CNN=CNN:cuda()
However, the speed seems to be unchanged when using cuda and without cuda. I do not know why?