How to programmatically generate deploy.prototxt in python

[Deepak Roy Chittajallu]

I have written python code to programmatically generate a convolutional neural network (CNN) for training and validation using caffe. Below is my function:

def custom_net(lmdb, batch_size):

    # define your own net!
    n = caffe.NetSpec()

    # keep this data layer for all networks
    n.data, n.label = L.Data(batch_size=batch_size, backend=P.Data.LMDB, source=lmdb,
                             ntop=2, transform_param=dict(scale=1. / 255))

    n.conv1 = L.Convolution(n.data, kernel_size=6,
                            num_output=48, weight_filler=dict(type='xavier'))
    n.pool1 = L.Pooling(n.conv1, kernel_size=2, stride=2, pool=P.Pooling.MAX)

    n.conv2 = L.Convolution(n.pool1, kernel_size=5,
                            num_output=48, weight_filler=dict(type='xavier'))
    n.pool2 = L.Pooling(n.conv2, kernel_size=2, stride=2, pool=P.Pooling.MAX)

    n.conv3 = L.Convolution(n.pool2, kernel_size=4,
                            num_output=48, weight_filler=dict(type='xavier'))
    n.pool3 = L.Pooling(n.conv3, kernel_size=2, stride=2, pool=P.Pooling.MAX)

    n.conv4 = L.Convolution(n.pool3, kernel_size=2,
                            num_output=48, weight_filler=dict(type='xavier'))
    n.pool4 = L.Pooling(n.conv4, kernel_size=2, stride=2, pool=P.Pooling.MAX)

    n.fc1 = L.InnerProduct(n.pool4, num_output=50,
                           weight_filler=dict(type='xavier'))

    n.drop1 = L.Dropout(n.fc1, dropout_param=dict(dropout_ratio=0.5))

    n.score = L.InnerProduct(n.drop1, num_output=2,
                             weight_filler=dict(type='xavier'))

    # keep this loss layer for all networks
    n.loss = L.SoftmaxWithLoss(n.score, n.label)

    return n.to_proto()

with open('net_train.prototxt', 'w') as f:
    f.write(str(custom_net(train_lmdb_path, train_batch_size)))

with open('net_test.prototxt', 'w') as f:
    f.write(str(custom_net(test_lmdb_path, test_batch_size)))

Similarly, is there a way to generate the deploy.prototxt in python?

If so, can someone point me to a reference.

I found this article -- https://chmlngm.wordpress.com/2015/09/06/hello-world/ -- but it looks very hackish. Wondering if there is a cleaner way?