def patch_ceilometer():
"""Add patches which are not merged to upstream
Order of patches applied:
ceilometer-poll-cpu-util.patch
ceilometer-rates-always-zero.patch
ceilometer-support-network-bytes.patch
"""
patchfile_list = [
'ceilometer-poll-cpu-util.patch',
'ceilometer-rates-always-zero.patch',
'ceilometer-support-network-bytes.patch',
]
for patch_file in patchfile_list:
utils.patch(DIST_PACKAGES_DIR, patch_file, 1)
def patch_nova_conductor():
"""Add patches which are not merged to upstream
Order of patches applied:
live-migration-vifmapping-controller.patch
"""
patchfile_list = [
# Change-Id: If0fb5d764011521916fbbe15224f524a220052f3
'live-migration-vifmapping-controller.patch',
]
for patch_file in patchfile_list:
utils.patch(utils.DIST_PACKAGES_DIR, patch_file, 1)
# Restart related service
utils.execute('service', 'nova-conductor', 'restart')
def patch_ceilometer():
"""Add patches which are not merged to upstream
Order of patches applied:
ceilometer-poll-cpu-util.patch
ceilometer-rates-always-zero.patch
ceilometer-support-network-bytes.patch
"""
patchfile_list = [
'ceilometer-poll-cpu-util.patch',
'ceilometer-rates-always-zero.patch',
'ceilometer-support-network-bytes.patch',
]
for patch_file in patchfile_list:
utils.patch(DIST_PACKAGES_DIR, patch_file, 1)
def createdData(X, label):
for c in range(OUTPUT_CLASSES):
PATCH, LABEL, TEST_PATCH, TRAIN_PATCH, TEST_LABEL, TRAIN_LABEL = [], [], [], [], [], []
for h in range(X.shape[1] - PATCH_SIZE + 1):
for w in range(X.shape[2] - PATCH_SIZE + 1):
gt = label[h, w]
if (gt == c + 1):
img = patch(X, PATCH_SIZE, h, w)
PATCH.append(img)
LABEL.append(gt - 1)
# 打乱切片
shuffle(PATCH)
# 划分测试集与训练集
split_size = int(len(PATCH) * TEST_FRAC)
TEST_PATCH.extend(PATCH[:split_size]) # 0 ~ split_size
TRAIN_PATCH.extend(PATCH[split_size:]) # split_size ~ len(class)
TEST_LABEL.extend(LABEL[:split_size])
TRAIN_LABEL.extend(LABEL[split_size:])
# 写入文件夹
train_dict, test_dict = {}, {}
train_dict["train_patches"] = TRAIN_PATCH
train_dict["train_labels"] = TRAIN_LABEL
file_name = "Training_class(%d).mat" % c
scipy.io.savemat(os.path.join(NEW_DATA_PATH, file_name), train_dict)
test_dict["testing_patches"] = TEST_PATCH
test_dict["testing_labels"] = TEST_LABEL
file_name = "Testing_class(%d).mat" % c
scipy.io.savemat(os.path.join(NEW_DATA_PATH, file_name), test_dict)
def patch_compute_xenapi():
"""Add patches which are not merged to upstream
Order of patches applied:
support-disable-image-cache.patch
speed-up-config-drive.patch
ovs-interim-bridge.patch
neutron-security-group.patch
"""
patchfile_list = [
'support-disable-image-cache.patch',
'speed-up-config-drive.patch',
'ovs-interim-bridge.patch',
'neutron-security-group.patch',
]
for patch_file in patchfile_list:
utils.patch(DIST_PACKAGES_DIR, patch_file, 1)
def patch_compute_xenapi():
"""Add patches which are not merged to upstream
Order of patches applied:
support-disable-image-cache.patch
speed-up-config-drive.patch
ovs-interim-bridge.patch
neutron-security-group.patch
"""
patchfile_list = [
'support-disable-image-cache.patch',
'speed-up-config-drive.patch',
'ovs-interim-bridge.patch',
'neutron-security-group.patch',
]
for patch_file in patchfile_list:
utils.patch(DIST_PACKAGES_DIR, patch_file, 1)
def patch_compute_xenapi():
"""Add patches which are not merged to upstream
Order of patches applied:
support-disable-image-cache.patch
speed-up-config-drive.patch
ovs-interim-bridge.patch
neutron-security-group.patch
live-migration-iscsi.patch
support-vif-hotplug.patch
fix-rescue-vm.patch
live-migration-vifmapping.patch
"""
patchfile_list = [
# Change-Id: I5ebff2c1f7534b06233a4d41d7f5f2e5e3b60b5a
'support-disable-image-cache.patch',
# Change-Id: I359e17d6d5838f4028df0bd47e4825de420eb383
'speed-up-config-drive.patch',
# Change-Id: I0cfc0284e1fcd1a6169d31a7ad410716037e5cc2
'ovs-interim-bridge.patch',
# Change-Id: Id9b39aa86558a9f7099caedabd2d517bf8ad3d68
'neutron-security-group.patch',
# Change-Id: I88d1d384ab7587c428e517d184258bb517dfb4ab
'live-migration-iscsi.patch',
# Change-Id: I22f3fe52d07100592015007653c7f8c47c25d22c
'support-vif-hotplug.patch',
# Change-Id: I32c66733330bc9877caea7e2a2290c02b3906708
'fix-rescue-vm.patch',
# Change-Id: If0fb5d764011521916fbbe15224f524a220052f3
'live-migration-vifmapping.patch',
# TODO(huanxie): below patch isn't merged into upstream yet,
# it only affects XS7.1 and later
# Change-Id: I31850b25e2f32eb65a00fbb824b08646c9ed340a
'assert_can_migrated.patch',
]
for patch_file in patchfile_list:
utils.patch(utils.DIST_PACKAGES_DIR, patch_file, 1)
def patch_neutron_ovs_agent():
"""Apply neutron patch
Add conntrack-tools patch to support conntrack in Dom0
"""
utils.patch('/usr/bin', 'fix-xenapi-returncode.patch', 2)
def patch_neutron_ovs_agent():
"""Apply neutron patch
Add conntrack-tools patch to support conntrack in Dom0
"""
utils.patch('/usr/bin', 'fix-xenapi-returncode.patch', 2)
import matplotlib.pyplot as plt
import pickle
import os
from nda import log
from nda.problems import LogisticRegression
from nda.optimizers import *
from nda.optimizers.utils import generate_mixing_matrix
from nda.experiment_utils import run_exp
import tikzplotlib
import time
from utils import patch
patch()
def plot_exp(exps, configs, filename, dim, n_agent, logx=False, logy=False):
colors = ['k', 'r', 'g', 'b', 'c', 'm', 'y']
line_styles = ['-', '--', ':']
# log.info("Initial accuracy = " + str(p.accuracy(x_0)))
results = [[exp.get_name()] + list(exp.get_metrics()) for exp in exps]
with open(f"data/{filename}", 'wb') as f:
pickle.dump(results, f)
row_index = {'t': 1, 'comm_rounds': 0}
column_index = {'f': 0}
outputs = np.zeros((data.shape[1], (data.shape[2]))) # 创建等尺寸的 0矩阵接收预测值
data = standartize(data) # 标准化
data = pad(data, 8) # 填充边框 与切片填充大小一致
# -----------生成网络模型---------------
logdir = "07-05_09-42-34" # 日志文件
NET_PARAMS_PATH = os.path.join(os.getcwd(), "log", logdir, "net_params.pkl")
net_params = torch.load(NET_PARAMS_PATH) # 加载训练好的模型参数
cnn = Net() # 生成网络
cnn.load_state_dict(net_params) # 参数放进网络
if torch.cuda.is_available(): # 使用GPU
cnn = cnn.cuda()
# -----------将数据输入模型-------------
for h in range(data.shape[1] - slice_size + 1):
for w in range(data.shape[2] - slice_size + 1):
img = patch(data, slice_size, h, w)
img = np.expand_dims(img, axis=0) # 再增加一个维度（模型接受三维数据）
img = torch.from_numpy(img).float()
if torch.cuda.is_available():
img = img.cuda()
output = cnn(img)
_, predicted = torch.max(output.data, 1)
outputs[h, w] = np.array(predicted.cpu())[0]
# -----------将target保存为图片------------
pic_name = logdir + ".tif"
SAVE_PATH = os.path.join(os.getcwd(), "predicted", pic_name)
PIL.Image.fromarray(outputs).save(SAVE_PATH)
# PIL.Image.fromarray(target_mat).save('D:/Code/test.tif')