**Source Link**

https://github.com/DeepScale/SqueezeNet

**Download Link** 

https://github.com/DeepScale/SqueezeNet/blob/master/SqueezeNet_v1.1/squeezenet_v1.1.caffemodel
https://github.com/DeepScale/SqueezeNet/blob/master/SqueezeNet_v1.1/train_val.prototxt


**Project Page** 

https://github.com/DeepScale/SqueezeNet

**Authors** 

Original Paper Title: SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size
Authors: Forrest N. Iandola and Song Han and Matthew W. Moskewicz and Khalid Ashraf and William J.

Caffe Implementation: http://deepscale.ai

**Labels**

Imagenet Labels
https://gist.github.com/yrevar/942d3a0ac09ec9e5eb3a

**License**

BSD License

Redistribution and use in source and binary forms, with or without modification, are permitted
provided that the following conditions are met:

1. Redistributions of source code must retain the above copyright notice, this list of conditions
and the following disclaimer.

2. Redistributions in binary form must reproduce the above copyright notice, this list of conditions
and the following disclaimer in the documentation and/or other materials provided with the
distribution.

THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR
IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND
FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR
CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER
IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF
THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

https://github.com/DeepScale/SqueezeNet/blob/master/LICENSE

**Changes Required**

The train_valid.prototxt needs to be updated for deployment. To do so, the following changes are required:

1. The data layer should be replaced by an input layer with dimensions equal to the crop size in the data layer.
2. The accuracy layers from the end of the prototxt file need to be deleted as they are not required during
   deployment. They can remain in the prototxt but they will be ignored.
3. The loss layer will be ignored, and a layer of type softmax is required to be added, with bottom layer as
   pool10 and top as "prob"
4. In the conversion call, the biases provided in the train_valid data layer are required to be given as arguments