discount=0.99,
target_freq=10,
verbose=True,
print_every=10)
'''
agent = DQNAgent(action_set=[0, 1, 2],
reward_function=mountain_car_reward_function,
feature_extractor=MountainCarIdentityFeature(),
hidden_dims=[50, 50],
learning_rate=5e-4,
buffer_size=50000,
batch_size=64,
num_batches=100,
starts_learning=5000,
final_epsilon=0.02,
discount=0.99,
target_freq=10,
verbose=True,
print_every=10)
_, _, rewards = live(agent=agent,
environment=env,
num_episodes=episodes,
max_timesteps=200,
verbose=True,
print_every=50)
np.save(os.path.join(reward_path, file_name), rewards)
agent.save(path=os.path.join(agent_path, file_name + '.pt'))
def __init__(self, options):
super(atropine, self).__init__()
self.vchannels = collections.OrderedDict()
self.fullscreen = options.fullscreen
self.no_escape = options.no_escape
self.max_brightness = options.max_brightness
gm = guide_manager.guide_manager(self, options)
cm = callsign_manager.callsign_manager(self, options)
im = icon_manager.icon_manager(cm, options)
gm.new_guide.connect(cm.new_guide)
gm.new_guide.connect(self.guide_update)
self.vchannel = None
self.source = None
self.resume_source = None
self.channel_file = options.channel_file
self.video = video_vlc.video_vlc()
self.sources = dict()
if len(options.hdhr_lineup_id):
for lineup, tuners in options.hdhr_lineup_id.iteritems():
self.sources[lineup] = source_hdhr.source_hdhr(
self.video, tuners)
else:
self.sources[''] = source_hdhr.source_hdhr(self.video)
self.live = live.live(self, self.video, im, self.vchannels)
self.guide = guide.guide_widget(self, options, self.video, im,
self.vchannels)
self.blank = Qt.QWidget(self)
self.addWidget(self.live)
self.addWidget(self.guide)
self.addWidget(self.blank)
self.video.setParent(self)
self.video.show()
self.resume_widget = None
self.setCurrentWidget(self.live)
self.live.clicked.connect(
lambda w=self.guide: self.setCurrentWidget(w))
self.guide.done.connect(lambda w=self.live: self.setCurrentWidget(w))
self.paused_start = options.paused
Qt.QTimer.singleShot(0, self.start)
ss = """
QWidget {
border: none;
background-color: rgb(60, 75, 90);
}
QLabel {
font-family: sans-serif;
font: 24pt;
color: white;
}
video, video_proxy, guide_widget now_widget, time_header_widget {
background-color: black
}
guide_widget station_logo_widget {
background-color: rgb(150, 170, 190);
}
guide_widget program_info_widget {
background-color: rgb(30, 30, 30);
border: 2px solid black;
}
guide_widget program_info_widget QWidget {
background-color: rgb(30, 30, 30);
}
guide_header_widget {
background-color: darkblue;
}
station_info_guide_large QWidget {
background-color: rgb(90, 105, 120);
}
info_widget, info_widget QWidget {
background: none;
}
program_label {
padding: 4px;
border: 2px solid black;
}
program_label:focus {
border: 6px solid white;
}
time_header_widget, now_widget {
padding: 8px;
}
station_info_widget {
padding: 6px;
}
osd_widget {
margin: 50px;
}
guide_widget > station_info_widget {
padding: 4px;
font-size: 20pt;
}
QProgressBar {
border: none;
color: red;
}
QProgressBar::chunk {
background-color: red;
}
"""
# https://raw.githubusercontent.com/MythTV/mythtv/master/mythtv/themes/default/categories.xml
cat_file = os.path.join(__location__, 'categories.xml')
cc = categories.category_colors(cat_file)
for key, value in cc.iteritems():
c = Qt.QColor(*value)
c = Qt.QColor.fromHsvF(c.hsvHueF(), c.hsvSaturationF(),
min(c.valueF(), self.max_brightness))
#c = Qt.QColor.fromHslF(c.hslHueF(), c.hslSaturationF(), min(c.lightnessF(), self.max_brightness))
c = str(c.getRgb()[0:3])
ss += 'QLabel[category="%s"] { background-color: rgb%s; }\n' % (
key, c)
self.setStyleSheet(ss)
from sklearn.naive_bayes import BernoulliNB
from sklearn.tree import DecisionTreeClassifier
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
from sklearn.pipeline import Pipeline, FeatureUnion
from sklearn.metrics import accuracy_score
from sklearn.metrics import confusion_matrix
from sklearn import metrics
import warnings
import math
from sklearn import preprocessing
p = "./models/shape_predictor_68_face_landmarks.dat"
detector = dlib.get_frontal_face_detector()
predictor = dlib.shape_predictor(p)
from calibration import calibration
from live import live
#Run Calibration
mean, std = calibration()
print(mean, std)
data, result = live(mean, std)
for D, r in zip(data, result):
print(D)
print(r)
print('\n')
channel = client.get_channel(840211266869919784)
await channel.send(f"{member} has left the server.")
@tasks.loop(seconds=300)
async def count():
temp = client.get_channel(839976544692731985)
channel = client.get_channel(840442030785691678)
await channel.edit(name="Server Users: " + str(temp.guild.member_count))
@client.command()
async def clear(ctx, amount=sys.maxsize - 1):
if (amount != sys.maxsize - 1):
await ctx.channel.purge(limit=amount + 1)
else:
await ctx.channel.purge(limit=amount)
@client.command()
async def test(ctx):
await ctx.send(ctx.guild.member_count)
# End of Bot
live()
client.run(token)
action_set=[0, 1, 2],
# reward_function=functools.partial(cartpole_reward_function, reward_type='sparse'),
reward_function=functools.partial(Forex_reward_function),
feature_extractor=ForexIdentityFeature(),
hidden_dims=[50, 50],
learning_rate=5e-4,
buffer_size=5000,
# batch_size=16,
batch_size=8,
num_batches=100,
starts_learning=100,
final_epsilon=0.02,
discount=0.99,
target_freq=10,
verbose=False,
print_every=10)
_, _, rewards = live(
agent=agent,
environment=env,
# num_episodes=100,
# max_timesteps=3601,
num_episodes=500,
max_timesteps=3600,
verbose=True,
print_every=50)
file_name = '|'.join(['dqn', str(seed)])
np.save(os.path.join(reward_path, file_name), rewards)
agent.save(path=os.path.join(agent_path, file_name + '.pt'))