This fork of Caffe adds the capability to dynamically mask
at a random position within training images, a mask_size
block, once in mask_freq
iterations.
It is motivated by the discussion on Kaggle forum
for State-farm distracted driver detection competition, where in it has been observed that
dynamically masking a small random portion of the images during training will:
- Augment training images at run-time, allowing training to proceed without saturation for more epochs
- Force the network to isolate and learn the contribution of each feature to objective
See data layer tutorial for more details.
Caffe is a deep learning framework made with expression, speed, and modularity in mind. It is developed by the Berkeley Vision and Learning Center (BVLC) and community contributors.
Check out the project site for all the details like
- DIY Deep Learning for Vision with Caffe
- Tutorial Documentation
- BVLC reference models and the community model zoo
- Installation instructions
and step-by-step examples.
Please join the caffe-users group or gitter chat to ask questions and talk about methods and models. Framework development discussions and thorough bug reports are collected on Issues.
Happy brewing!
Caffe is released under the BSD 2-Clause license. The BVLC reference models are released for unrestricted use.
Please cite Caffe in your publications if it helps your research:
@article{jia2014caffe,
Author = {Jia, Yangqing and Shelhamer, Evan and Donahue, Jeff and Karayev, Sergey and Long, Jonathan and Girshick, Ross and Guadarrama, Sergio and Darrell, Trevor},
Journal = {arXiv preprint arXiv:1408.5093},
Title = {Caffe: Convolutional Architecture for Fast Feature Embedding},
Year = {2014}
}