Caffe

    Check out our web image classification demo!

    Expressive architecture encourages application and innovation.Models and optimization are defined by configuration without hard-coding.Switch between CPU and GPU by setting a single flag to train on a GPU machine then deploy to commodity clusters or mobile devices.

    Extensible code fosters active development.In Caffe’s first year, it has been forked by over 1,000 developers and had many significant changes contributed back.Thanks to these contributors the framework tracks the state-of-the-art in both code and models.

    Speed makes Caffe perfect for research experiments and industry deployment.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.

    Community: 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 and Github.

    • With the ILSVRC2012-winning model and prefetching IO.
    • DIY Deep Learning for Vision with Caffe and Tutorial presentation of the framework and a full-day crash course.
    • Tutorial DocumentationPractical guide and framework reference.
    • Tested on Ubuntu, Red Hat, OS X.
    • Model ZooBAIR suggests a standard distribution format for Caffe models, and provides trained models.
    • Guidelines for development and contributing to Caffe.
    • API DocumentationDeveloper documentation automagically generated from code comments.
    • Comparison of inference and learning for different networks and GPUs.

    • Image Classification and Filter VisualizationInstant recognition with a pre-trained model and a tour of the net interface for visualizing features and parameters layer-by-layer.

    • Define, train, and test the classic LeNet with the Python interface.

    • Fine-tuning for Style RecognitionFine-tune the ImageNet-trained CaffeNet on new data.

    • Use Caffe as a generic SGD optimizer to train logistic regression on non-image HDF5 data.

    • R-CNN detectionRun a pretrained model as a detector in Python.

    • Extracting features and plotting the Siamese network embedding.

    Command Line Examples

    • Train and test "CaffeNet" on ImageNet data.

    • LeNet MNIST TutorialTrain and test "LeNet" on the MNIST handwritten digit data.

    • Train and test Caffe on CIFAR-10 data.

    • Fine-tuning for style recognitionFine-tune the ImageNet-trained CaffeNet on the "Flickr Style" dataset.

    • Extract CaffeNet / AlexNet features using the Caffe utility.

    • CaffeNet C++ Classification exampleA simple example performing image classification using the low-level C++ API.

    • Train and test a siamese network on MNIST data.

    Please cite Caffe in your publications if it helps your research:

    If you do publish a paper where Caffe helped your research, we encourage you to cite the framework for tracking by Google Scholar.

    Join the to ask questions and discuss methods and models. This is where we talk about usage, installation, and applications.

    Framework development discussions and thorough bug reports are collected on Issues.

    The BAIR Caffe developers would like to thank NVIDIA for GPU donation, A9 and Amazon Web Services for a research grant in support of Caffe development and reproducible research in deep learning, and BAIR PI for guidance.

    The BAIR members who have contributed to Caffe are (alphabetical by first name):Carl Doersch, , Evan Shelhamer, , Jon Long, , Ronghang Hu, , Sergey Karayev, , Takuya Narihira, and .

    The open-source community plays an important and growing role in Caffe’s development.Check out the Github project pulse for recent activity and the for the full list.

    We sincerely appreciate your interest and contributions!If you’d like to contribute, please read the developing & contributing guide.