I am trying to use Brightway's ParallelMonteCarloand MultiMonteCarloclass but have run into a KeyError. I am in a Brightway project with an LCI database:
In [1] bw.databases
Out [1] Brightway2 databases metadata with 2 objects:
biosphere3
ecoinvent 3_2 CutOff
Selecting an activity and a method:
In [2] db = bw.Database('ecoinvent 3_2 CutOff')
act = db.random()
method = ('CML 2001', 'climate change', 'GWP 100a')
My code is as follows:
In [3] ParallelMC_LCA = bw.ParallelMonteCarlo({act:1},
method = myMethod,
iterations=1000,
cpus=mp.cpu_count())
results = np.array(ParallelMC_LCA.calculate())
and
In [4] act1 = db.random()
act2 = db.random()
multiMC_LCA = bw.MultiMonteCarlo(demands = [{act1:1}, {act2:1}],
method = myMethod,
iterations = 10)
results = np.array(ParallelMC_LCA.calculate())
Both give me a KeyError: 'ecoinvent 3_2 CutOff'.
My question is: why?
This is a known issue due to differences in how multiprocessing works on Windows and Unix. Specifically, on Windows the project is not set correctly, causing a KeyError. As such, it isn't a Stack Overflow question.
Related
I'm trying to implement Double DQN (not to be confused with DQN with a slightly delayed Q-target network) in PyTorch to train an agent to play an Atari OpenAI Gym game. Here I discuss the implementation of the following formula:
Update of Q-network, formula taken from Sutton & Barto.
My first implementation is:
Q_pred = self.Q_1.forward(s_now)[T.arange(batch_size), actions.long()]
Q_next_all = self.Q_1.forward(s_next)
maxA_id = T.argmax(Q_next_all, dim=1)
Q_pred2 = self.Q_2.forward(s_next)[T.arange(batch_size), maxA_id]
Q_target = (rewards + (~dones) * self.GAMMA * Q_pred2).detach()
self.Q_1.optimizer.zero_grad()
self.Q_1.loss(Q_target, Q_pred).backward()
self.Q_1.optimizer.step()
(Q_1 and Q_2 are nn.Module classes, and all of the variables involved here are already torch tensors lying in the GPU.)
I noticed that my program ran much slower than a previous implementation which used plain DQN.
I realized that I can combine the batches entering Q_1, so there will be one combined batch being forwarded in the neural network, instead of two batches in sequence. The code becomes:
s_combined = T.cat((s_now, s_next))
Q_combined = self.Q_1.forward(s_combined)
Q_pred = Q_combined[T.arange(batch_size), actions.long()]
Q_next_all = Q_combined[batch_size:]
Q_pred2_all = self.Q_2.forward(s_next)
maxA_id = T.argmax(Q_next_all, dim=1)
Q_pred2 = Q_pred2_all[T.arange(batch_size), maxA_id]
Q_target = (rewards + (~dones) * self.GAMMA * Q_pred2).detach()
self.Q_1.optimizer.zero_grad()
self.Q_1.loss(Q_target, Q_pred).backward()
self.Q_1.optimizer.step()
(This proves that I understand how to do batch training in PyTorch, so don't mark this as a duplicate of this question.)
Furthermore, I realized that Q_1 and Q_2 can process their batches in parallel. So I looked up how to do multiprocessing in PyTorch. Unfortunately, I couldn't find a good example. I tried to adapt a code that looks similar to my scenario, and my code becomes:
def spawned():
s_combined = T.cat((s_now, s_next))
Q_combined = self.Q_1.forward(s_combined)
Q_pred = Q_combined[T.arange(batch_size), actions.long()]
Q_next_all = Q_combined[batch_size:]
mp.set_start_method('spawn', force=True)
p = mp.Process(target=spawned)
p.start()
Q_pred2_all = self.Q_2.forward(s_next)
p.join()
maxA_id = T.argmax(Q_next_all, dim=1)
Q_pred2 = Q_pred2_all[T.arange(batch_size), maxA_id]
Q_target = (rewards + (~dones) * self.GAMMA * Q_pred2).detach()
self.Q_1.optimizer.zero_grad()
self.Q_1.loss(Q_target, Q_pred).backward()
self.Q_1.optimizer.step()
This crashes with the error message:
AttributeError: Can't pickle local object 'Agent.learn.<locals>.spawned'
So how do I make this work?
(Achieving this in CUDA programming is trivial. One simply launches two device kernels using a sequential host code, and the two kernels are automatically computed in parallel in the GPU.)
I have recently started exploring the QuantLib option pricing libraries for python and have come across an error that I don't seem to understand. Basically, I am trying to price an Up&Out Barrier option using the Heston model. The code that I have written has been taken from examples found online and adapted to my specific case. Essentially, the problem is that when I run the code below I get an error that I believe is triggered at the last line of the code, i.e. the european_option.NPV() function
*** RuntimeError: wrong argument type
Can someone please explain me what I am doing wrong?
# option inputs
maturity_date = ql.Date(30, 6, 2020)
spot_price = 969.74
strike_price = 1000
volatility = 0.20
dividend_rate = 0.0
option_type = ql.Option.Call
risk_free_rate = 0.0016
day_count = ql.Actual365Fixed()
calculation_date = ql.Date(26, 6, 2020)
ql.Settings.instance().evaluationDate = calculation_date
# construct the option payoff
european_option = ql.BarrierOption(ql.Barrier.UpOut, Barrier, Rebate,
ql.PlainVanillaPayoff(option_type, strike_price),
ql.EuropeanExercise(maturity_date))
# set the Heston parameters
v0 = volatility*volatility # spot variance
kappa = 0.1
theta = v0
hsigma = 0.1
rho = -0.75
spot_handle = ql.QuoteHandle(ql.SimpleQuote(spot_price))
# construct the Heston process
flat_ts = ql.YieldTermStructureHandle(ql.FlatForward(calculation_date,
risk_free_rate, day_count))
dividend_yield = ql.YieldTermStructureHandle(ql.FlatForward(calculation_date,
dividend_rate, day_count))
heston_process = ql.HestonProcess(flat_ts, dividend_yield,
spot_handle, v0, kappa,
theta, hsigma, rho)
# run the pricing engine
engine = ql.AnalyticHestonEngine(ql.HestonModel(heston_process),0.01, 1000)
european_option.setPricingEngine(engine)
h_price = european_option.NPV()
The problem is that the AnalyticHestonEngine is not able to price Barrier options.
Check here https://www.quantlib.org/reference/group__barrierengines.html for a list of Barrier Option pricing engines.
I am in trouble in simulating the EnergyPlus-FMU by PyFMI. I created an EnergyPlus FMU using the reference building model. I am using PyFMI2.5. How do I run the do_step() function?
from pyfmi import load_fmu
model = load_fmu("MyEnergyplus.fmu")
start_time = 0
final_time = 60.0 * 60 * 24 * 3 #seconds
step_size = 60 # seconds
opts = model.simulate_options()
idf_steps_per_hour = 60
ncp = (final_time - start_time)/(3600./idf_steps_per_hour)
opts['ncp'] = ncp
t = 0
status = model.do_step(current_t = t, step_size= step_size, new_step=True)
The error I got:
File "test_fmi2.py", line 15, in <module> status = model.do_step(current_t = t, step_size= step_size, new_step=True)
AttributeError: 'pyfmi.fmi.FMUModelME2' object has no attribute 'do_step'
I double checked the APIs of PyFMI, and didn't find any problem.
How to enable the simulation? Thanks.
From the output we can see that the FMU that you have loaded is an Model Exchange FMU that do not have a do step function (only Co-Simulation FMUs have that). For more information about the different FMU types, please see the FMI specification.
To simulate an Model Exchange FMU, please use the "simulate" method. The "simulate" method is also available for Co-Simulation FMUs and is the prefered way to perform a simulation
Not knowing how you setup the fmu, I can at least say that you forgot model.initialize(start_time,final_time).
I'm trying to connect two blocks with the "Connector" class implemented in pyomo, using the following simple examplary code.
from pyomo.environ import *
m = ConcreteModel()
# Block 01
m.block_01 = Block()
m.block_01.flow = Var(within=NonNegativeReals, bounds=(2, 10))
m.block_01.OUT = Connector(initialize= {'flow': m.block_01.flow})
# Block 02
m.block_02 = Block()
m.block_02.flow = Var(within=NonNegativeReals)
m.block_02.IN = Connector(initialize= {'flow': m.block_02.flow})
m.con = Constraint(expr=m.block_01.OUT == m.block_02.IN)
def _obj(_m):
return _m.block_01.flow + _m.block_02.flow
m.obj = Objective(rule=_obj)
After "optimization" all variables take their lower bound values (m.block_01.flow = 2 and m.block_02.flow = 0). So the Connector seems not to transfer any data for the variables.
If I'm using:
m.con = Constraint(expr=m.block_01.flow == m.block_02.flow)
instead, it works. However this is not the idea of Connectors, right?
Any ideas about the reason for the problem?
Did you apply the expand_connectors transformation before sending your model to a solver?
TransformationFactory('core.expand_connectors').apply_to(m)
I am trying to use tf.train.batch to run enqueue images in multiple threads. When the number of threads is 1, the code works fine. But when I set a higher number of threads I receive an error:
Failed precondition: Attempting to use uninitialized value Variable
[[Node: Variable/read = Identity[T=DT_INT32, _class=["loc:#Variable"], _device="/job:localhost/replica:0/task:0/cpu:0"](Variable)]]
The main thread has to run for some time under one second to index the database of folders and put it into tensor.
I tried to use sess.run([some_image]) before running tf.train.bath loop. In that case workers fail in the background first with the same error, and after that I receive my images.
I tried to use time.sleep(), but it does not seem to be possible to delay the workers.
I tried adding a dependency to the batch:
g = tf.get_default_graph()
with g.control_dependencies([init_one,init_two]):
example_batch = tf.train.batch([my_image])
where init_one, and init_two are tf.initialize_all(variables) and tf.initialize_local_variables()
the most relevant issue I could find is at: https://github.com/openai/universe-starter-agent/issues/44
Is there a way I could ask the synchronize worker threads with the main thread so that they don't race first and die out ?
A similar and easy to reproduce error with variable initialization happens when set up the epoch counter to anything that is not None Are there any potential solutions ? I've added the code needed to reproduce the error below:
def index_the_database(database_path):
"""indexes av4 database and returns two tensors of filesystem path: ligand files, and protein files"""
ligand_file_list = []
receptor_file_list = []
for ligand_file in glob(os.path.join(database_path, "*_ligand.av4")):
receptor_file = "/".join(ligand_file.split("/")[:-1]) + "/" + ligand_file.split("/")[-1][:4] + '.av4'
if os.path.exists(receptor_file):
ligand_file_list.append(ligand_file)
receptor_file_list.append(receptor_file)
index_list = range(len(ligand_file_list))
return index_list,ligand_file_list, receptor_file_list
index_list,ligand_file_list,receptor_file_list = index_the_database(database_path)
ligand_files = tf.convert_to_tensor(ligand_file_list,dtype=tf.string)
receptor_files = tf.convert_to_tensor(receptor_file_list,dtype=tf.string)
filename_queue = tf.train.slice_input_producer([ligand_files,receptor_files],num_epochs=10,shuffle=True)
serialized_ligand = tf.read_file(filename_queue[0])
serialized_receptor = tf.read_file(filename_queue[1])
image_one = tf.reduce_sum(tf.exp(tf.decode_raw(serialized_receptor,tf.float32)))
image_batch = tf.train.batch([image_one],100,num_threads=100)
init_two = tf.initialize_all_variables()
init_one = tf.initialize_local_variables()
sess = tf.Session()
coord = tf.train.Coordinator()
threads = tf.train.start_queue_runners(sess=sess,coord=coord)
sess.run([init_one])
sess.run([init_two])
while True:
print "next"
sess.run([image_batch])