wenguanwang / deepattention Goto Github PK

Hi,

Could you tell me which model generated the precomputed maps? Is it the model trained on salicon or fine-tuned on other datasets?

Thanks.

Hi!
Please can you say which software you used to draw your Fig. 3. Architecture of our deep attention model?
And if there is any pytorch implementation of your paper to recommend.

Thank you.

Since the provided caffe is an old one, it can't be compiled when cudnn>=5.0 and opencv<3.0. So you are supposed to comment USE_CUDNN := 1 when compiling caffe and make sure your opencv works.

Hi, @wenguanwang ,

Could you share your train_prototxt file? Thanks!

Hi, could you provide the training code?
Thanks.