原书出处
Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)
Backpropagation does not distinguish between parameters and data — it treats both as generic inputs to parameterless modules. Therefore, we can use backprop to optimize data inputs to the graph just like we can use backprop to optimize parameter inputs to the graph.
原书定位详情
来源:a19c7933-4eda-4176-aebd-30fe7e1ad292
正文位置:extracted/Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)/_epub-src/OPS/xhtml/c014.xhtml
EPUB SHA-256:60a7eebc43f5d9a675d776fed3736fbdaa791e55877023f2f8cba016a9b384f8