Weight update of the 2-input model

jongyookim / IQA_DeepQA_FR_release

CNN-based full-reference image quality assessment

64 stars 23 forks source link

Weight update of the 2-input model #2

Open Catheriana opened 6 years ago

Catheriana commented 6 years ago

Thanks again for your code. I am interested in the 2-input(distorted and error map) model. And I wonder if some changes in the process of back propogating the weight are needed after we concatenate the output of the Conv2-1 and Conv2-2. Thanks a lot!