Awesome

Caffe-compact

Caffe-compact aims to provide a self-contained CNN model testing library.

This project remove most unnecessary dependency for CNN net testing and feature extraction. Note that we completely remove CUDA dependency in caffe-compact.

Current dependency:

• c++11 compiler (for shared_ptr)
• google protobuf

Optional dependency:

• cblas (e.g. libatlas3gf-base)
• Eigen3

You can select an matrix backend by setting the USE_EIGEN environment in the Makefile.

These dependencies can be satisfied on most platform including Windows and mobile. It makes Caffe-compact much easier to deploy.

This work also avoids potential license problems along with the third-party libraris when release your caffe CNN model.

Difference

The original project can be found at: https://github.com/BVLC/caffe Caffe-compact only support a subset of functionality of caffe:

• CNN forward pass only
• CPU only
• Raw image input only

Performance

MKL has performance problem when dealing with small matrix (e.g. testing your model on only one input image), especially multithreading is enabled. Atlas or other open source BLAS implementation may perform better.

TODO: benchmark

Future Work

• integrate protobuf

Yuheng Chen, 2014 chyh1990@gmail.com