Significance of roi_heads.box_head() features

[shubham-scisar]

Can anyone please guide that after I load the model and proposals into the roi_heads.box_head,I get a tensor of 1000 proposals ,each proposal is a tensor of 1024 features which are float values ,please guide as to what these features signify and is there any other way that these features can be visualised on the image or the intermediate feature maps obtained from the p2,p3,p4,p5,p6 layers. Please guide.