def page_fault(algorithm, pages, frames, results):
if algorithm not in results:
if algorithm == "FIFO":
results[algorithm] = fifo(pages, frames)
elif algorithm == "LRU":
results[algorithm] = lru(pages, frames)
elif algorithm == "LFU":
results[algorithm] = lfu(pages, frames)
elif algorithm == "OPT":
results[algorithm] = opt(pages, frames)
elif algorithm == "RAND":
results[algorithm] = rand(pages, frames)
elif algorithm == "MFU":
results[algorithm] = mfu(pages, frames)
elif algorithm == "MRU":
results[algorithm] = mru(pages, frames)
else:
return -1
return results[algorithm]
def ILP_optA_solution(tasks, cpu, tk, model=Model()):
# init
A = cpu.A
M = cpu.M
Sk = tk.s
tasks = [t for t in tasks if t != tk]
n = len(tasks)
I = [t.W(Sk) for t in tasks]
a = [t.a for t in tasks]
Ak = opt(tasks, cpu, tk)
if (n <= 0):
return 0
if (Ak == 0):
Ak = A
# Ak = A
# model = Model()
model.clear()
x = {
i: model.add_var(obj=0, var_type="C", name="x[%d]" % i)
for i in range(n)
}
y = {
i: model.add_var(obj=0, var_type="C", name="y[%d]" % i)
for i in range(n)
}
model.objective = maximize(
xsum((1 / M) * x[i] + (a[i] / Ak) * y[i] for i in range(n)))
model.verbose = False
for i in range(n):
model += x[i] + y[i] <= I[i]
for i in range(n):
model += x[i] <= xsum((1 / M) * x[j] for j in range(n))
for i in range(n):
model += y[i] <= xsum((a[j] / Ak) * y[j] for j in range(n))
model.optimize()
return model.objective_value
def main(testfunc, Npop=100, Ngen=100, repeat=1, Extra=None, chromname='chrom'):
"""
testfunc --> name of MOOP test function
Ngenome --> number of total possible genes (# pressure ports, infinite for math)
constr --> list of gene numbers that are constrained (not used)
repeat --> number of duplcate runs to do
Extra --> extra text to add to filename
chromname --> name of chromosome, can be specific to optimization
"""
#MAKE OPTIMIZATION OBJECT
optobj = opt(testfunc, dup=repeat)
#keys of objective function (traits we are optimizing form)
params = optobj.optfunc.params
#number of genes in chromosome (number of independent variables)
ngenes = optobj.optfunc.ngenes
#MAKE GENETIC ALGORITHM OBJECT
#Initialize NSGA2 object
gaobj = nsga2(optobj, params, Npop=Npop, Ngen=Ngen)
#Run evolution for Ngen generations
gaobj.Evolution()
gaobj.SaveData(chromname)
print('\n\nBest Individual:')
print(gaobj.best['chrom'])
#Print members of final pareto frontier
print("\n\nFinal Pareto Frontier Individuals' Chromosomes:")
for f in gaobj.mate['fronts'][0]:
print(gaobj.mate['indv'][f]['chrom'])
num_epochs = 100
save_every = 50
print_every = 10
valid_every = 20 # test the valid data when batch step is (int)
grad_clip = 5.
learning_rate = 0.001
# Store every options to opt class data structure
opt = opt(data_dir=data_dir,
save_dir=save_dir,
gen_dir=gen_dir,
ratio=ratio,
num_layers=num_layers,
hidden_size=hidden_size,
embedding_size=embedding_size,
cuda=cuda,
batch_size=batch_size,
seq_len=seq_len,
num_epochs=num_epochs,
save_every=save_every,
print_every=print_every,
valid_every=valid_every,
grad_clip=grad_clip,
learning_rate=learning_rate)
# load the vocab data
with open('vocab/vocab.pkl', 'rb') as f:
vocab = pickle.load(f)
# load the hidden data if resume is true
if args.resume:
with open(os.path.join(opt.gen_dir, "hidden.pkl"), 'rb') as f:
def uphkidxoptall():
hkequ.hkequ()
idx.idx()
opt.opt()
def uphkidxoptall(): hkequ.hkequ() idx.idx() opt.opt()
import os
import sys
import csv
import itemSimList
import filterCustRate
import opt
def recommendItem(rateList,rec_itemList):
rec_userItem={}
rec_items=[]
for user,itemList in rateList.items():
if user not in rec_userItem:
rec_userItem[user]=[]
for items in itemList:
rec_items=rec_itemList[items]
for rec_item in rec_items:
rec_userItem[user].append(rec_item)
return rec_userItem
if __name__=="__main__":
#sys.argv[1] should be opt file name
rateList=filterCustRate.filterCustRate("custRating.csv")
rec_itemList=itemSimList.itemSimList("itemSim.csv")
rec_userItem=recommendItem(rateList,rec_itemList)
#print rec_userItem
opt.opt(rec_userItem,sys.argv[1])
print("Optimisation has run.")
else:
if fg == True:
print("Set first_guess to True")
input()
opt.opt_fg()
print("Optimisation has run.")
print("Run experiment, press any button")
print("Set first_guess to False")
input()
for i in range(50):
os.system('cnee --replay -sp 15 -f xnee.xns >/dev/null 2>&1')
time.sleep(10)
opt.opt()
print("Replication is finished.")
sys.exit()
clean = False
fi_gu = False
while clean == False:
print("Do you want to start a new experiment?")
print("Old data will be moved and you have to restart.")
print("Confirm with yes.")
aux = input()
aux = aux.lower()
if (aux in ['no', 'nein', 'n'] or aux[0] in ['n', 'm']):
print("You don't want to restart.")
clean = True
import pylab from numpy import * from numpy.linalg import * from opt import opt import file_io path = "reg_test/" restart = "restart.out" niter = 1 # max ifDual iterations # read in initial turbulent viscosity temp = path temp += "nut_no.dat" nno, x_no, y_no, nu0 = file_io.read_field(temp) # create an instance of opt o = opt(x_no, y_no, niter, restart, path) tol = 1e-9 nopt = 200 # number of optimization iterations x_opt, J_opt, res = o.lbfgs(nopt, nu0, tol) nu_opt = exp(x_opt) print J_opt # write out mu_opt temp = path temp += "nu_opt" file_io.write_field(temp, x_no, y_no, nu_opt)
from numpy import * from numpy.linalg import * from opt import opt import file_io path = "c004/" restart = "restart.out" niter = 50000 # max ifDual iterations np = 6 # number of cores to run on ctolpri = 2e-6 ctoladj = 1e-4 # read in initial turbulent viscosity temp = path; temp+="nut_no.dat" nno, x_no, y_no, nu0 = file_io.read_field(temp) # create an instance of opt o = opt(x_no, y_no, niter, restart, path, np, ctolpri, ctoladj) tol = 1e-9 nopt = 500 # number of optimization iterations x_opt, J_opt, res = o.lbfgs(nopt, nu0, tol) nu_opt = exp(x_opt) print J_opt # write out mu_opt temp = path; temp+="nut_opt" file_io.write_field(temp, x_no, y_no, nu_opt)
from numpy.linalg import * from opt import opt import file_io path = "c004/" restart = "restart.out" niter = 50000 # max ifDual iterations np = 6 # number of cores to run on ctolpri = 2e-6 ctoladj = 1e-4 # read in initial turbulent viscosity temp = path temp += "nut_no.dat" nno, x_no, y_no, nu0 = file_io.read_field(temp) # create an instance of opt o = opt(x_no, y_no, niter, restart, path, np, ctolpri, ctoladj) tol = 1e-9 nopt = 500 # number of optimization iterations x_opt, J_opt, res = o.lbfgs(nopt, nu0, tol) nu_opt = exp(x_opt) print J_opt # write out mu_opt temp = path temp += "nut_opt" file_io.write_field(temp, x_no, y_no, nu_opt)
for i in range(len(timestep) - 2))
lstm = []
lstm.append(
seg_LSTM(input_size,
hidden_size=hidden_size,
num_layers=num_layers,
bi_directional=bi_directional) for i in range(len(len) - 1))
return lstm
def get_parameters(*args):
return nn.Parameter()
if __name__ == '__main__':
opts = opt.opt()
scale = [4, 8, 16]
work_directory = opts.directory
annotation_directory = os.path.join(work_directory, opts.annotation)
output_path = os.path.join(work_directory, opts.output)
resnext_model = generate_model(opts)
conv_block = []
conv_block.append(scale_block(_, 1024, opts.cardinality) for _ in scale)
LSTM_part = LSTM(1024, opts.time_steps, opts.num_layers, opts.bi)
parameters = get_parameters()
optimizer = optim.SGD(parameters,
lr=0.1,
momentum=0.9,
dampening=opts.dampening,
nesterov=opts.nesterov)
