Instant Recognition with Caffe

In this example we'll classify an image with the bundled CaffeNet model based on the network architecture of Krizhevsky et al. for ImageNet. We'll compare CPU and GPU operation then reach into the model to inspect features and the output.

(These feature visualizations follow the DeCAF visualizations originally by Yangqing Jia.)

Based on: https://github.com/BVLC/caffe/blob/master/examples/00-classification.ipynb

First, import required modules, set plotting parameters, and run ./scripts/download_model_binary.py models/bvlc_reference_caffenet to get the pretrained CaffeNet model if it hasn't already been fetched.


In [1]:
import numpy as np
import matplotlib.pyplot as plt
%matplotlib inline

# Make sure that caffe is on the python path:
caffe_root = '/home/ubuntu/digits/caffe/'  # this file is expected to be in {caffe_root}/examples
import sys
sys.path.insert(0, caffe_root + 'python')

import caffe

plt.rcParams['figure.figsize'] = (10, 10)
plt.rcParams['image.interpolation'] = 'nearest'
plt.rcParams['image.cmap'] = 'gray'

import os
if not os.path.isfile(caffe_root + 'models/bvlc_reference_caffenet/bvlc_reference_caffenet.caffemodel'):
    print("Downloading pre-trained CaffeNet model...")
#    !../scripts/download_model_binary.py ../models/bvlc_reference_caffenet

Set Caffe to CPU mode, load the net in the test phase for inference, and configure input preprocessing.


In [2]:
caffe.set_mode_cpu()
net = caffe.Net(caffe_root + 'models/bvlc_reference_caffenet/deploy.prototxt',
                caffe_root + 'models/bvlc_reference_caffenet/bvlc_reference_caffenet.caffemodel',
                caffe.TEST)

# input preprocessing: 'data' is the name of the input blob == net.inputs[0]
transformer = caffe.io.Transformer({'data': net.blobs['data'].data.shape})
transformer.set_transpose('data', (2,0,1))
transformer.set_mean('data', np.load(caffe_root + 'python/caffe/imagenet/ilsvrc_2012_mean.npy').mean(1).mean(1)) # mean pixel
transformer.set_raw_scale('data', 255)  # the reference model operates on images in [0,255] range instead of [0,1]
transformer.set_channel_swap('data', (2,1,0))  # the reference model has channels in BGR order instead of RGB

In [3]:
mean = np.load(caffe_root + 'python/caffe/imagenet/ilsvrc_2012_mean.npy')
print mean.shape
print mean.mean(1).mean(1)


(3, 256, 256)
[ 104.00698793  116.66876762  122.67891434]

Let's start with a simple classification. We'll set a batch of 50 to demonstrate batch processing, even though we'll only be classifying one image. (Note that the batch size can also be changed on-the-fly.)


In [4]:
# set net to batch size of 50
net.blobs['data'].reshape(50,3,227,227)

Feed in the image (with some preprocessing) and classify with a forward pass.


In [5]:
net.blobs['data'].data[...] = transformer.preprocess('data', caffe.io.load_image(caffe_root + 'examples/images/fish-bike.jpg'))
out = net.forward()
print("Predicted class is #{}.".format(out['prob'].argmax()))


Predicted class is #770.

What did the input look like?


In [6]:
plt.imshow(transformer.deprocess('data', net.blobs['data'].data[0]))


Out[6]:
<matplotlib.image.AxesImage at 0x7f5d890d1a90>

Adorable, but was our classification correct?


In [8]:
# load labels
imagenet_labels_filename = caffe_root + 'data/ilsvrc12/synset_words.txt'
try:
    labels = np.loadtxt(imagenet_labels_filename, str, delimiter='\t')
except:
    !../data/ilsvrc12/get_ilsvrc_aux.sh
    labels = np.loadtxt(imagenet_labels_filename, str, delimiter='\t')

# sort top k predictions from softmax output
top_k = net.blobs['prob'].data[0].flatten().argsort()[-1:-6:-1]
print labels[top_k]


['n04120489 running shoe'
 'n02835271 bicycle-built-for-two, tandem bicycle, tandem'
 'n04482393 tricycle, trike, velocipede' 'n04509417 unicycle, monocycle'
 'n03125729 cradle']

Indeed! But how long did it take?


In [9]:
# CPU mode
net.forward()  # call once for allocation
%timeit net.forward()


1 loop, best of 3: 2.53 s per loop

That's a while, even for a batch size of 50 images. Let's switch to GPU mode.


In [10]:
# GPU mode
caffe.set_device(0)
caffe.set_mode_gpu()
net.forward()  # call once for allocation
%timeit net.forward()


1 loop, best of 3: 93.4 ms per loop

Much better. Now let's look at the net in more detail.

First, the layer features and their shapes (1 is the batch size, corresponding to the single input image in this example).


In [32]:
[(k, v.data.shape) for k, v in net.blobs.items()]


Out[32]:
[('data', (50, 3, 227, 227)),
 ('conv1', (50, 96, 55, 55)),
 ('pool1', (50, 96, 27, 27)),
 ('norm1', (50, 96, 27, 27)),
 ('conv2', (50, 256, 27, 27)),
 ('pool2', (50, 256, 13, 13)),
 ('norm2', (50, 256, 13, 13)),
 ('conv3', (50, 384, 13, 13)),
 ('conv4', (50, 384, 13, 13)),
 ('conv5', (50, 256, 13, 13)),
 ('pool5', (50, 256, 6, 6)),
 ('fc6', (50, 4096)),
 ('fc7', (50, 4096)),
 ('fc8', (50, 1000)),
 ('prob', (50, 1000))]

The parameters and their shapes. The parameters are net.params['name'][0] while biases are net.params['name'][1].


In [33]:
[(k, v[0].data.shape) for k, v in net.params.items()]


Out[33]:
[('conv1', (96, 3, 11, 11)),
 ('conv2', (256, 48, 5, 5)),
 ('conv3', (384, 256, 3, 3)),
 ('conv4', (384, 192, 3, 3)),
 ('conv5', (256, 192, 3, 3)),
 ('fc6', (4096, 9216)),
 ('fc7', (4096, 4096)),
 ('fc8', (1000, 4096))]

Helper functions for visualization


In [34]:
# take an array of shape (n, height, width) or (n, height, width, channels)
# and visualize each (height, width) thing in a grid of size approx. sqrt(n) by sqrt(n)
def vis_square(data, padsize=1, padval=0):
    data -= data.min()
    data /= data.max()
    
    # force the number of filters to be square
    n = int(np.ceil(np.sqrt(data.shape[0])))
    padding = ((0, n ** 2 - data.shape[0]), (0, padsize), (0, padsize)) + ((0, 0),) * (data.ndim - 3)
    data = np.pad(data, padding, mode='constant', constant_values=(padval, padval))
    
    # tile the filters into an image
    data = data.reshape((n, n) + data.shape[1:]).transpose((0, 2, 1, 3) + tuple(range(4, data.ndim + 1)))
    data = data.reshape((n * data.shape[1], n * data.shape[3]) + data.shape[4:])
    print data.shape
    plt.imshow(data)

The input image

The first layer filters, conv1


In [35]:
# the parameters are a list of [weights, biases]
filters = net.params['conv1'][0].data
vis_square(filters.transpose(0, 2, 3, 1))


(120, 120, 3)

The first layer output, conv1 (rectified responses of the filters above, first 36 only)


In [30]:
feat = net.blobs['conv1'].data[0, :36]
vis_square(feat, padval=1)


The second layer filters, conv2

There are 256 filters, each of which has dimension 5 x 5 x 48. We show only the first 48 filters, with each channel shown separately, so that each filter is a row.


In [31]:
filters = net.params['conv2'][0].data
vis_square(filters[:48].reshape(48**2, 5, 5))


The second layer output, conv2 (rectified, only the first 36 of 256 channels)


In [32]:
feat = net.blobs['conv2'].data[0, :36]
vis_square(feat, padval=1)


The third layer output, conv3 (rectified, all 384 channels)


In [33]:
feat = net.blobs['conv3'].data[0]
vis_square(feat, padval=0.5)


The fourth layer output, conv4 (rectified, all 384 channels)


In [34]:
feat = net.blobs['conv4'].data[0]
vis_square(feat, padval=0.5)


The fifth layer output, conv5 (rectified, all 256 channels)


In [35]:
feat = net.blobs['conv5'].data[0]
vis_square(feat, padval=0.5)


The fifth layer after pooling, pool5


In [36]:
feat = net.blobs['pool5'].data[0]
vis_square(feat, padval=1)


The first fully connected layer, fc6 (rectified)

We show the output values and the histogram of the positive values


In [37]:
feat = net.blobs['fc6'].data[0]
plt.subplot(2, 1, 1)
plt.plot(feat.flat)
plt.subplot(2, 1, 2)
_ = plt.hist(feat.flat[feat.flat > 0], bins=100)


The second fully connected layer, fc7 (rectified)


In [38]:
feat = net.blobs['fc7'].data[0]
plt.subplot(2, 1, 1)
plt.plot(feat.flat)
plt.subplot(2, 1, 2)
_ = plt.hist(feat.flat[feat.flat > 0], bins=100)


The final probability output, prob


In [39]:
feat = net.blobs['prob'].data[0]
plt.plot(feat.flat)


Out[39]:
[<matplotlib.lines.Line2D at 0x7f4e804d5850>]

Let's see the top 5 predicted labels.


In [25]:
# load labels
imagenet_labels_filename = caffe_root + 'data/ilsvrc12/synset_words.txt'
try:
    labels = np.loadtxt(imagenet_labels_filename, str, delimiter='\t')
except:
    !../data/ilsvrc12/get_ilsvrc_aux.sh
    labels = np.loadtxt(imagenet_labels_filename, str, delimiter='\t')

# sort top k predictions from softmax output
top_k = net.blobs['prob'].data[0].flatten().argsort()[-1:-6:-1]
print labels[top_k]


['n02123045 tabby, tabby cat' 'n02123159 tiger cat'
 'n02124075 Egyptian cat' 'n02119022 red fox, Vulpes vulpes'
 'n02127052 lynx, catamount']

In [ ]: