ACKNOWLEDGMENTS

Some of the components of this project are derived/inspired from/by the following sources:

1.  Google's CNN benchmark project which is hosted here:
      https://github.com/tensorflow/benchmarks/tree/master/scripts/tf_cnn_benchmarks
    For more details, see this web site:
      https://www.tensorflow.org/performance
    Along with this benchmarking suite, we distribute one of the older versions of this
    project.

2.  Some of the Caffe2's benchmarking project components are derived/inspired by
    Caffe2's examples on GitHub:
    https://github.com/caffe2/caffe2/

3.  Some of the MXNet's benchmarking project components are derived/inspired by
    MXNet's examples on GitHub:
    https://github.com/apache/incubator-mxnet/blob/master/example/image-classification/

4.  Some of the Caffe's prototxt models (AlexNet and GoogleNet) are taken from
    BVLC Caffe's web site:
    https://github.com/BVLC/caffe/tree/master/models

5.  Some of the Caffe's ResNet models are from here:
    https://github.com/KaimingHe/deep-residual-networks

6.  Caffe's VGG16 and VGG19 models are taken from here:
    https://gist.github.com/ksimonyan/211839e770f7b538e2d8#file-readme-md
    https://gist.github.com/ksimonyan/3785162f95cd2d5fee77#file-readme-md

7.  Frameworks' docker files are based on docker files from respective repositories:
      BVLC Caffe:    https://github.com/BVLC/caffe/tree/master/docker/gpu
      Caffe2:        https://github.com/caffe2/caffe2/tree/master/docker
      Intel Caffe:   https://github.com/intel/caffe/tree/master/docker
      MXNet:         https://github.com/apache/incubator-mxnet/tree/master/docker
      NVIDIA Caffe:  https://github.com/NVIDIA/caffe/tree/caffe-0.16/docker
      TensorFlow:    https://github.com/tensorflow/tensorflow/tree/master/tensorflow/tools/docker

8.  Bash command line parser is the one used in Kaldi project:
    https://github.com/kaldi-asr/kaldi/blob/master/egs/wsj/s5/utils/parse_options.sh

9.  To visualize neural network architectures, we use Netscope project:
    https://github.com/ethereon/netscope

10. Documentation is served with docsify:
    https://github.com/QingWei-Li/docsify/

11. TensorRT benchmark C++ code is based on NVIDIA's `giexec.cpp` sample shipped with
    TensorRT distribution package.

12. ResNet18 and ResNet34 inference/training Caffe's prototxt descriptors are based
    on this Anting Shen's git hub repository (https://github.com/antingshen/resnet-protofiles).
    They match reference implementations from Kaiming He's repository (see #5 above).

13. Caffe's InceptionV3 descriptors are based on this deployment config:
    https://github.com/soeaver/caffe-model/blob/master/cls/inception/deploy_inception-v3.prototxt

14. Caffe's InceptionV4 descriptors are based on this deployment config:
    https://github.com/soeaver/caffe-model/blob/master/cls/inception/deploy_inception-v4.prototxt

15. A large number of optimization tricks have been borrowed from NVIDIA's docker containers from
    NGC.

16. Contributing guidelines are based on guidelines from POSH-HPOneView project:
    https://github.com/HewlettPackard/POSH-HPOneView