zhreshold

"""References:
Andrew G. Howard, Menglong Zhu, Bo Chen, Dmitry Kalenichenko, Weijun Wang,
Tobias Weyand, Marco Andreetto, Hartwig Adam.
"MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications."
arXiv preprint arXiv:1704.04861
"""
import mxnet as mx

def depthwise_conv(data, kernel, pad, num_filter, num_group, stride, name):
    conv = mx.symbol.Convolution(data=data, kernel=kernel, pad=pad, stride=stride,

zhreshold

import mxnet as mx

def inflated_layer(data, num_in, num_out, name):
    assert(num_out % num_in == 0)
    num_group = num_out / num_in
    outputs = []
    for i in range(num_group):
        bias = mx.sym.Variable(shape=(1, num_in, 1, 1),
            name="{}_{}_bias".format(name, i))
        outputs.append(mx.sym.broadcast_add(lhs=data, rhs=bias))

zhreshold

"""References:

Simonyan, Karen, and Andrew Zisserman. "Very deep convolutional networks for
large-scale image recognition." arXiv preprint arXiv:1409.1556 (2014).
"""
import mxnet as mx

def depthwise_conv(data, kernel, pad, num_filter, name, num_group):
    conv = mx.symbol.Convolution(data=data, kernel=kernel, pad=pad,
        num_filter=num_group, name=name+'_depthwise', num_group=num_group)

zhreshold

import numpy as np

OLD_CLASSES = ['aeroplane', 'bicycle', 'bird', 'boat', 'bottle', 'bus', 'car',
               'cat', 'chair', 'cow', 'diningtable', 'dog', 'horse', 'motorbike',
               'person', 'pottedplant', 'sheep', 'sofa', 'train', 'tvmonitor']

NEW_CLASSES = ['bicycle', 'bus', 'car', 'cat', 'cow', 'dog', 'horse', 'motorbike',
               'person', 'train']

OLD_INDICES = [OLD_CLASSES.index(x) for x in NEW_CLASSES]

zhreshold

Techniques to accelerate convolutional neural networks

• Depthwise convolution (mobilenet)
• Shufflenet
• Densenet

zhreshold

import mxnet as mx
print(mx.__version__)
import time
import os

print('worker:', os.environ.get('MXNET_CPU_WORKER_NTHREADS', 1))

num_batch = 100
batch_size = 32
data_shape = (3, 224, 224)

zhreshold

name: "GoogleNet"
layer {
  name: "data"
  type: "Input"
  top: "data"
  input_param { shape: { dim: 10 dim: 3 dim: 224 dim: 224 } }
}
layer {
  name: "conv1/7x7_s2"
  type: "Convolution"

zhreshold

import mxnet as mx
import numpy as np
from timeit import default_timer as timer

def get_bench_net(num_hidden=10000):
  data = mx.sym.var('data')
  fc = mx.sym.FullyConnected(data, num_hidden=num_hidden)
  return fc

num_out = 10000

zhreshold

python train_imagenet.py --data-train ~/data/train.rec --data-val ~/data/val.rec --gpus 0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15 --data-nthreads 32 --network resnet101_v2 --batch-size 256 --top-k 5 --model-prefix model/resnet101_v2 --min-random-scale 0.533 --max-random-shear-ratio 0 --max-random-rotate-angle 0 --max-random-h 0 --max-random-l 0 --max-random-s 0 --lr-step-epochs 30,60,90 --num-epochs 120 --rgb-std '58.395,57.12,57.375'

zhreshold

(function() {

// Chart design based on the recommendations of Stephen Few. Implementation
// based on the work of Clint Ivy, Jamie Love, and Jason Davies.
// http://projects.instantcognition.com/protovis/bulletchart/
d3.bullet = function() {
  var orient = "left", // TODO top & bottom
      reverse = false,
      duration = 0,
      ranges = bulletRanges,