def dostuff(num):
start = time.Time()
x=0
for x in range(num):
x = x+1
end = time.Time()
print(totaltime = end - start)
return x
def check_for_surge(self, bytes_received, packets_received):
if bytes_received > SwitchStats.surge_threshold / SwitchStats.polling_interval:
end_timestamp = time.Time()
surge_data = {
'datapath': self.datapath,
'start_timestamp':
end_timestamp - SwitchStats.polling_interval,
'end_timestamp': end_timestamp,
'packets_received': packets_received,
'bytes_received': bytes_received
}
if self.current_surge != None:
self.current_surge.extend(surge_data)
else:
self.current_surge = TrafficSurge(surge_data)
return self.current_surge
else:
if self.current_surge != None:
self.current_surge.end()
surge = copy.copy(self.current_surge)
self.current_surge = None
return surge
else:
return None
def json_detection():
"""
Returns JSON with classes found in the images
"""
raw_images = []
images = request.files.getlist("images")
for image in images:
image_name = image.filename
image_names.append(image_name)
image.save(os.path.join(os.getcwd(), image_name))
img_raw = tf.image.decode_image(
open(image_name, 'rb').read(), channels=3
)
raw_images.append(img_raw)
num = 0
# list for final response
for i in range(len(raw_images)):
'''
list of responses for current image
'''
response = []
responses = []
raw_img = raw_images[i]
num += 1
img = tf.expand_dims(raw_img, 0)
img = transform_images(img, size)
time1 = time.Time()
boxes, scores, classes, num = yolo(img)
time2 = time.time()
print("Time: {}".format(time2 - time1))
print("Detections:")
for j in range(nums[0]):
print("\t{}, {}, {}".format(class_names[int(classes[0][j])],
np.array(scores[0][j]),
np.array(boxes[0][j])))
resposes.append([
"class": class_names[int(classes[0][j])],
"confidence": float("{o:.2f}".format(np.array(scores[0][j]*100)))
])
response.append([
"image": image_names[j],
"detections": responses
])
def Scan(a, m, n):
time = Time()
count = 0
for i in range(m, m + 64 + 1):
for j in range(n, n + 64 + 1):
if a[i, j] >= 200:
count += 1
if count >= 64 * 64 * 1 / 2:
time.Time()
if time.Check() and not x[-1] == True and not x[-2] == True:
# return True
x.append(True)
else:
x.append(False)
else:
x.append(False)
def handleRequest(req):
env = pack.unpackRequest(req)
if env is False:
return False
head = env.Head
cur_time = head.Time
appkey = head.AppKey[:8]
#
# 通过appkey可以在游戏表里面获取appsecret
#
appsecret = appkey
secret = binascii.unhexlify(appsecret)
if head.Platform != common.PT_ANDROID and head.Platform != common.PT_IOS:
env.Head.Result = common.MP_ERR_PLATFORM
return pack.packRequest(env, "")
key = pack.GetKey(secret, cur_time, head.Platform)
result = ymcoder.mpDecode(env.ReqBody, key)
if result is False:
env.Head.Result = common.MP_ERR_DECODE
return pack.packRequest(env, "")
body = zlib.decompress(result)
if body is False:
env.Head.Result = common.MP_ERR_UNPACK
return pack.packRequest(env, "")
env.HeadExt.readFrom(body)
env.ReqBody = body[env.HeadExt.Len:]
retenv = handleProcess(env)
retenv.Head.Time = int(time.Time())
key = pack.GetKey(secret, retenv.Head.Time, head.Platform)
return pack.packRequest(env, key)
if len(contours) != 0:
#Draw in blue the found contours
cv2.drawContours(output, contours, -1, 255, 3)
#Find the largest contour
c = max(contours, key=cv2.contourArea)
#print str(cv2.contourArea(c))
#Contour must be this big to count as ball. If number too small when no ball present may detect anything
if cv2.contourArea(c) > 525:
#Draw contour with circle
(x, y), radius = cv2.minEnclosingCircle(c)
center = (int(x), int(y))
diameter = radius * 2
radius = int(radius)
cv2.circle(output, center, radius, (0, 255, 0), 2)
#Draw contour with a rectangle
'''
x, y, w, h = cv2.boundingRect(c)
#Draw contour in green
cv2.rectangle(output, (x,y),(x+w, y+h), (0, 255, 0),2)
'''
#Show the images
endTime = time.Time()
print "Time taken: " + str(endTime - startTime)
cv2.imshow("Ball", np.hstack([pImg, output]))
cv2.waitKey(0) # OpenCV for Linux has a bug and needs this line
cv2.destroyAllWindows()
def set_expstart(table):
from astropy import time
mjd = time.Time(table['t_min'], format='mjd')
table['expstart'] = mjd.iso
def train(train_loader, model, criterion, optimizer, epoch):
batch_time = AverageMeter()
data_time = AverageMeter()
losses = AverageMeter()
top1 = AverageMeter()
top5 = AverageMeter()
if args.no_partialbn:
model.module.partialBN(False)
else:
model.module.partialBN(True)
# 切换到训练模式
# 注意：调用的model重写的train方法以达到冻结bn层参数的目的
model.train()
end = time.Time()
for i, (input, target) in enumerate(train_loader):
# 计算数据的导入时间
# 当执行enumerate(train_loader)的时候，是先调用DataLoader类的__iter__方法，该方法里面再调用DataLoaderIter类的初始化操作__init__
# 而当执行for循环操作时，调用DataLoaderIter类的__next__方法，在该方法中通过self.collate_fn接口读取self.dataset数据时就会调用TSNDataSet类的__getitem__方法，从而完成数据的迭代读取
data_time.update(time.time() - end)
# 读取到数据后就将数据从Tensor转换成Variable格式，然后执行模型的前向计算
# 如果想在CUDA上进行计算，需要将操作对象放在GPU内存中。
target = target.cuda(async=True)
input_var = torch.autograd.Variable(input)
target_var = torch.autograd.Variable(target)
# 计算输出
# output:batch size*class维度
# 自动调用模型中自定义的forward函数
output = model(input_var)
loss = criterion(output, target_var)
# 计算准确度和视频的总损失
prec1, prec5 = accuracy(output.data, target, topk=(1, 5))
losses.update(loss.data[0], input.size(0))
prec1.update(prec1.data[0], input.size(0))
prec5.update(prec5.data[0], input.size(0))
# 对所有的参数的梯度缓冲区进行归零
optimizer.zero_grad()
# autograd.Variable 这是这个包中最核心的类。
# 它包装了一个Tensor，并且几乎支持所有的定义在其上的操作。一旦完成了你的运算，你可以调用 .backward()来自动计算出所有的梯度。
# 反向传播，自动计算梯度
loss.backward()
if args.clip_gradient is not None:
total_norm = clip_grad_norm(model.parameters(), args.clip_gradient)
if total_norm > args.clip_gradient:
print("clipping gradient: {} with coef {}".format(
total_norm, args.clip_gradient / total_norm))
# 执行参数更新
optimizer.step()
# 计算消耗的时间
batch_time.update(time.time() - end)
end = time.time()
if i % args.print_freq == 0:
print(('Epoch: [{0}][{1}/{2}], lr: {lr:.5f}\t'
'Time {batch_time.val:.3f} ({batch_time.avg:.3f})\t'
'Data {data_time.val:.3f} ({data_time.avg:.3f})\t'
'Loss {loss.val:.4f} ({loss.avg:.4f})\t'
'Prec@1 {top1.val:.3f} ({top1.avg:.3f})\t'
'Prec@5 {top5.val:.3f} ({top5.avg:.3f})'.format(
epoch,
i,
len(train_loader),
batch_time=batch_time,
data_time=data_time,
loss=losses,
top1=top1,
top5=top5,
lr=optimizer.param_groups[-1]['lr'])))
def __init__(self, surge_data):
super(EventTrafficSurgeEnd, self).__init__()
self.timestamp = time.Time()
self.surge_data = surge_data
def Updater(self):
self.nextTime = time.Time() - self.startTime
self.SetTime(self.nextTime)
self.timer = self.after(50, self.Updater)
def ping(self):
lc = time.Time()