A notebook to easily visualize filters, activations and output probability at each layer. This is extending caffe's classify.py; please refer to Caffe's Licence.
Please install Caffe, Python, Cuda, CUDNN. Download the models.
Setup the locations in the notebook for the models.
In the notebook, set a url for an image, hit the button: "Submit for run:". All the following cells will be computed.
The notebook will identify, "Convolution", "InnerProduct" layers and create dropdown boxes. Selecting any convolution layer will allow, visualizaion of that layer.