Why is torch.get_num_threads returning 1 despite setting it to NUM_THREADS =12

Question

I am very new to using torch and I am trying to run a python script using torch. The block code is as follows:

torch.backends.cudnn.deterministic = False
torch.backends.cudnn.benchmark = True
# uniform thread number
torch.set_num_threads(NUM_THREADS)
print('THREADS: ',torch.get_num_threads())
assert NUM_THREADS == torch.get_num_threads(), torch.get_num_threads()# Code fails here

Here is my.env file:

CPU_NUM_THREADS=12
OMP_NUM_THREADS=12
OPENMP_NUM_THREADS=12
OPENBLAS_NUM_THREADS=12
MKL_NUM_THREADS=12
VECLIB_MAXIMUM_THREADS=12
NUMEXPR_NUM_THREADS=12

When I tried printing NUM_THREADS, it outputs 12 but when I print torch.get_num_threads, it outputs 1.

My System info:

3.1 GHz 6-Core Intel Core i5 MacOS.

I would like to know why is torch.get_num_threads() outputting 1 instead of 12 and how can I fix it?

Answer 1

Update: I updated my torch version from 1.5.0 to 1.7.1 and that resolved the issue.