Speed + Simplicity
Caffe was built for performance, and when combined with Python, you get:
Fast execution (especially for CNNs)
Easy model prototyping using .prototxt files + Python scripting Review collected by and hosted on G2.com.
Outdated & Less Maintained
Lacking modern features (like dynamic computation graphs) Review collected by and hosted on G2.com.
Caffe is a deep learning framework, developed at the University of California, Berkeley, open-source, with a BSD license. It is written in C++, with a Python interface. Review collected by and hosted on G2.com.
It is a language with a slow execution speed. Review collected by and hosted on G2.com.
The best thing i like about caffee is the community contribution made by developers. Review collected by and hosted on G2.com.
The algo is still in development so cant say much about dislikes. Review collected by and hosted on G2.com.