Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Accelerating image processing for CNN #668

Merged
merged 8 commits into from
Dec 8, 2016
273 changes: 273 additions & 0 deletions python/paddle/utils/image_multiproc.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,273 @@
import os, sys
import numpy as np
from PIL import Image
from cStringIO import StringIO
import multiprocessing
from functools import partial

from paddle.utils.image_util import *
from paddle.trainer.config_parser import logger

Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

加上__all__字段,把需要Export的东西,Export出来。

Copy link
Collaborator

@reyoung reyoung Dec 5, 2016

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

加入 __all__,管理export的符号。

try:
import cv2
except ImportError:
logger.warning("OpenCV2 is not installed, using PIL to prcoess")
cv2 = None


class CvTransfomer(ImageTransformer):
"""
CvTransfomer used python-opencv to process image.
"""

def __init__(
self,
min_size=None,
crop_size=None,
transpose=(2, 0, 1), # transpose to C * H * W
channel_swap=None,
mean=None,
is_train=True,
is_color=True):
ImageTransformer.__init__(self, transpose, channel_swap, mean, is_color)
self.min_size = min_size
self.crop_size = crop_size
self.is_train = is_train

def resize(self, im, min_size):
row, col = im.shape[:2]
new_row, new_col = min_size, min_size
if row > col:
new_row = min_size * row / col
else:
new_col = min_size * col / row
im = cv2.resize(im, (new_row, new_col), interpolation=cv2.INTER_CUBIC)
return im

def crop_and_flip(self, im):
"""
Return cropped image.
The size of the cropped image is inner_size * inner_size.
im: (H x W x K) ndarrays
"""
row, col = im.shape[:2]
start_h, start_w = 0, 0
if self.is_train:
start_h = np.random.randint(0, row - self.crop_size + 1)
start_w = np.random.randint(0, col - self.crop_size + 1)
else:
start_h = (row - self.crop_size) / 2
start_w = (col - self.crop_size) / 2
end_h, end_w = start_h + self.crop_size, start_w + self.crop_size
if self.is_color:
im = im[start_h:end_h, start_w:end_w, :]
else:
im = im[start_h:end_h, start_w:end_w]
if (self.is_train) and (np.random.randint(2) == 0):
if self.is_color:
im = im[:, ::-1, :]
else:
im = im[:, ::-1]
return im

def transform(self, im):
im = self.resize(im, self.min_size)
im = self.crop_and_flip(im)
# transpose, swap channel, sub mean
im = im.astype('float32')
ImageTransformer.transformer(self, im)
return im

def load_image_from_string(self, data):
flag = cv2.CV_LOAD_IMAGE_COLOR if self.is_color else cv2.CV_LOAD_IMAGE_GRAYSCALE
im = cv2.imdecode(np.fromstring(data, np.uint8), flag)
return im

def transform_from_string(self, data):
im = self.load_image_from_string(data)
return self.transform(im)

def load_image_from_file(self, file):
flag = cv2.CV_LOAD_IMAGE_COLOR if self.is_color else cv2.CV_LOAD_IMAGE_GRAYSCALE
im = cv2.imread(file, flag)
return im

def transform_from_file(self, file):
im = self.load_image_from_file(file)
return self.transform(im)


class PILTransfomer(ImageTransformer):
"""
PILTransfomer used PIL to process image.
"""

def __init__(
self,
min_size=None,
crop_size=None,
transpose=(2, 0, 1), # transpose to C * H * W
channel_swap=None,
mean=None,
is_train=True,
is_color=True):
ImageTransformer.__init__(self, transpose, channel_swap, mean, is_color)
self.min_size = min_size
self.crop_size = crop_size
self.is_train = is_train

def resize(self, im, min_size):
row, col = im.size[:2]
new_row, new_col = min_size, min_size
if row > col:
new_row = min_size * row / col
else:
new_col = min_size * col / row
im = im.resize((new_row, new_col), Image.ANTIALIAS)
return im

def crop_and_flip(self, im):
"""
Return cropped image.
The size of the cropped image is inner_size * inner_size.
"""
row, col = im.size[:2]
start_h, start_w = 0, 0
if self.is_train:
start_h = np.random.randint(0, row - self.crop_size + 1)
start_w = np.random.randint(0, col - self.crop_size + 1)
else:
start_h = (row - self.crop_size) / 2
start_w = (col - self.crop_size) / 2
end_h, end_w = start_h + self.crop_size, start_w + self.crop_size
im = im.crop((start_h, start_w, end_h, end_w))
if (self.is_train) and (np.random.randint(2) == 0):
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

这里不需要加括号

im = im.transpose(Image.FLIP_LEFT_RIGHT)
return im

def transform(self, im):
im = self.resize(im, self.min_size)
im = self.crop_and_flip(im)
im = np.array(im, dtype=np.float32) # convert to numpy.array
# transpose, swap channel, sub mean
ImageTransformer.transformer(self, im)
return im

def load_image_from_string(self, data):
im = Image.open(StringIO(data))
return im

def transform_from_string(self, data):
im = self.load_image_from_string(data)
return self.transform(im)

def load_image_from_file(self, file):
im = Image.open(file)
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

最好不要使用file作为参数名,file在python里面是个函数

return im

def transform_from_file(self, file):
im = self.load_image_from_file(file)
return self.transform(im)


def warpper(cls, (dat, label)):
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

删了吧,没用

return cls.job(dat, label)


class MultiProcessImageTransformer(object):
def __init__(self,
procnum=10,
resize_size=None,
crop_size=None,
transpose=(2, 0, 1),
channel_swap=None,
mean=None,
is_train=True,
is_color=True,
is_img_string=True):
"""
Processing image with multi-process. If it is used in PyDataProvider,
the simple usage for CNN is as follows:

.. code-block:: python

def hool(settings, is_train, **kwargs):
settings.is_train = is_train
settings.mean_value = np.array([103.939,116.779,123.68], dtype=np.float32)
settings.input_types = [
dense_vector(3 * 224 * 224),
integer_value(1)]
settings.transformer = MultiProcessImageTransformer(
procnum=10,
resize_size=256,
crop_size=224,
transpose=(2, 0, 1),
mean=settings.mean_values,
is_train=settings.is_train)


@provider(init_hook=hook, pool_size=20480)
def process(settings, file_list):
with open(file_list, 'r') as fdata:
for line in fdata:
data_dic = np.load(line.strip()) # load the data batch pickled by Pickle.
data = data_dic['data']
labels = data_dic['label']
labels = np.array(labels, dtype=np.float32)
for im, lab in settings.dp.run(data, labels):
yield [im.astype('float32'), int(lab)]

:param procnum: processor number.
:type procnum: int
:param resize_size: the shorter edge size of image after resizing.
:type resize_size: int
:param crop_size: the croping size.
:type crop_size: int
:param transpose: the transpose order, Paddle only allow C * H * W order.
:type transpose: tuple or list
:param channel_swap: the channel swap order, RGB or BRG.
:type channel_swap: tuple or list
:param mean: the mean values of image, per-channel mean or element-wise mean.
:type mean: array, The dimension is 1 for per-channel mean.
The dimension is 3 for element-wise mean.
:param is_train: training peroid or testing peroid.
:type is_train: bool.
:param is_color: the image is color or gray.
:type is_color: bool.
:param is_img_string: The input can be the file name of image or image string.
:type is_img_string: bool.
"""

self.pool = multiprocessing.Pool(procnum)
self.is_img_string = is_img_string
if cv2 is not None:
self.transformer = CvTransfomer(resize_size, crop_size, transpose,
channel_swap, mean, is_train,
is_color)
else:
self.transformer = PILTransfomer(resize_size, crop_size, transpose,
channel_swap, mean, is_train,
is_color)

def run(self, data, label):
try:
fun = partial(warpper, self)
Copy link
Collaborator

@reyoung reyoung Dec 5, 2016

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

func = partial(MultiProcessImageTransformer. __job__, self.is_img_string, self.transformer)

Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

另外,最好不要 from functools import partial 而是 import functoolsfunctools.partial
这样可读性强一点点。

return self.pool.imap_unordered(fun, zip(data, label), chunksize=5)
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

有几个问题:

  • chunksize最好再设大一些,并且是procnum的倍数。例如 procnum * 100

    • 是procnum的倍数是因为这样各个进程间会比较均衡。
    • 再大一些是因为可以避免一些时候,某个核心被其他人占用,产生的抖动。
  • 推荐用itertools.izip替换zip。因为zip会先把list生成出来,然后再传递给imap_unordered函数,而izip会返回一个generator,一条一条的返回。这样首先内存占用应该理论上会小一些。同时,izip可以按照需要zip,应该会稍微快一些。

except KeyboardInterrupt:
self.pool.terminate()
except Exception, e:
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

这里不是阻塞的,所以这里try...except一点用应该都没有

self.pool.terminate()

Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@staticmethod
def __job__(is_img_string, transformer, data, label):
    if is_img_string:
       transformer. transform_from_string(data), label
    else:
       transformer. transform_from_file(data), label

def job(self, data, label):
if self.is_img_string:
return self.transformer.transform_from_string(data), label
else:
return self.transformer.transform_from_file(data), label

def __getstate__(self):
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

不要加这个函数了

self_dict = self.__dict__.copy()
del self_dict['pool']
return self_dict

def __setstate__(self, state):
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

不要加这个函数了

self.__dict__.update(state)
31 changes: 17 additions & 14 deletions python/paddle/utils/image_util.py
Original file line number Diff line number Diff line change
Expand Up @@ -186,29 +186,32 @@ def __init__(self,
channel_swap=None,
mean=None,
is_color=True):
self.transpose = transpose
self.channel_swap = None
self.mean = None
self.is_color = is_color
self.set_transpose(transpose)
self.set_channel_swap(channel_swap)
self.set_mean(mean)

def set_transpose(self, order):
if self.is_color:
assert 3 == len(order)
if order is not None:
if self.is_color:
assert 3 == len(order)
self.transpose = order

def set_channel_swap(self, order):
if self.is_color:
assert 3 == len(order)
if order is not None:
if self.is_color:
assert 3 == len(order)
self.channel_swap = order

def set_mean(self, mean):
# mean value, may be one value per channel
if mean.ndim == 1:
mean = mean[:, np.newaxis, np.newaxis]
else:
# elementwise mean
if self.is_color:
assert len(mean.shape) == 3
if mean is not None:
# mean value, may be one value per channel
if mean.ndim == 1:
mean = mean[:, np.newaxis, np.newaxis]
else:
# elementwise mean
if self.is_color:
assert len(mean.shape) == 3
self.mean = mean

def transformer(self, data):
Expand Down