Caffee

Caffee Use Cases

Caffe is a deep learning framework made with expression, speed, and modularity in mind. It is developed by Berkeley AI Research and by community contributors. Yangqing Jia created the project during his PhD at UC Berkeley. Caffe is released under the BSD 2-Clause license.

Caffe can process over 60M images per day with a single NVIDIA K40 GPU. That’s 1 ms/image for inference and 4 ms/image for learning and more recent library versions and hardware are faster still. We believe that Caffe is among the fastest convnet implementations available.

Caffe already powers academic research projects, startup prototypes, and even large-scale industrial applications in vision, speech, and multimedia. Join our community of brewers on the caffe-users group and Github.

Caffe provides extensive documentation including DIY Deep Learning for Vision with Caffe and Caffe in a Day tutorial presentations, tutorial documentation, arXiv / ACM MM ‘14 paper, installation instructions, model zoo, developing & contributing guidelines, API documentation, and benchmarking information.

Caffe also offers notebook examples for image classification and filter visualization, learning LeNet, fine-tuning for style recognition, off-the-shelf SGD for classification, multilabel classification with Python data layer, editing model parameters, R-CNN detection, and Siamese network embedding.

There are also command line examples for ImageNet tutorial, LeNet MNIST tutorial, CIFAR-10 tutorial, fine-tuning for style recognition, feature extraction with Caffe C++ code, CaffeNet C++ classification example, web demo, and Siamese Network tutorial.

Caffe can be cited in publications as a convolutional architecture for fast feature embedding. Contacting the Caffe team is possible through the caffe-users group and by submitting bug reports on Github.