Ying Tai（邰颖）
Ying Tai（邰颖）
Hi, I haven't tried our model with C++ interface. I'm afraid I can't give you any useful suggestions. Sorry for that.
Hi, which model are you using?
I think the difference comes from different frameworks we use. You can use torch to generate the LR image and see if the results are good or not.
Hi, Since I haven't tried this work on Ubuntu 18, I am not clear on the setup requirements steps either. Besides, I think our model can test with CPU, but...
Sorry, no docker image :(
I met the same problem when processing large amounts of images! I found it loaded some images twice, therefor some other images are skipped!
Thanks for your response. Actually, my question is what should I do to format those feature maps from the previous Conv layers to be the input of the ConvLSTM layer....
I see. Thanks! Another simple question is what is the effect of the second input "seq" containing 0 or 1 in this example?
I see. I have successfully trained your convLSTM version for a regression task. The training is good and the testing using command (e.g., $CAFFE test --model=XXX --weights=XXX) is also ok....
Hi, Here is some part of my prototxt, which concatenates 5 previous convs' outputs. The dimensions of kernel and input are 32x3x3 and 5x32x32x31x31, respectively. The batchsize is set to...