Hi, I tried to run multiprocessing on cpu for my network, but confused about the issue below:
import torch
import torch.multiprocessing as mp
import numpy as np
import time
params2 = ...
params5 = ...
images = ...
def run():
times = []
for i in [1,2, 3, 4]:
start = time.time()
mp.spawn(train, args=(images, params2, params5, 20, 10), nprocs=i)
#train(images, params2, params5, batch_size, epochs=1000)
end = time.time()
times.append(end - start)
np.savetxt('./times.txt', np.array(times))
if __name__ == '__main__':
run()
The error is TypeError: train() takes from 4 to 5 positional arguments but 6 were given. Actually, I provided 5 args, how it could be 6?