Jiaxiang Ren

From [these lines](https://github.com/abhaydoke09/Bilinear-CNN-TensorFlow/blob/cc2df0964c8d91523f68dc5f6c96fa7bd780f06b/core/bcnn_finetuning.py#L215-L230), '784' is the result of 28x28, which corresponds to 'feature dimension c' in Bilinear-CNN paper. Please refer to #7 and [these lines](https://github.com/abhaydoke09/Bilinear-CNN-TensorFlow/blob/cc2df0964c8d91523f68dc5f6c96fa7bd780f06b/utils/create_h5_dataset.py#L8-L14) for more info.

From this line: [Bilinear-CNN-TensorFlow/core/bcnn_finetuning.py#L226](https://github.com/abhaydoke09/Bilinear-CNN-TensorFlow/blob/cc2df0964c8d91523f68dc5f6c96fa7bd780f06b/core/bcnn_finetuning.py#L226) as for VGG, I think each "location" corresponds to each pixel in the last feature map. More generally, one location can correspond to one pixel or...

> Hi. It seems like the dataset that you use do not have a column `'response'`. You should check the name of columns of your dataset and those you want...