mseth10 commented on issue #14721: Random number generator seed setting does not always work for `mxnet.ndarray.linalg.potrf` URL: https://github.com/apache/incubator-mxnet/issues/14721#issuecomment-492876912 @iaroslav-ai , I tried reproducing the error, but wasn't able to. I ran the code given by you 10,000 times and it produces the same `L` every iteration. Also, I checked the code for `potrf` and found that the implementation does not involve a random generator, so having the same output is expected. I am using the same version of MXNet and ran the code on CPU. Are you using GPU? Code used (with slight modification): ``` import numpy as np import mxnet as mx from mxnet import nd K = nd.load("K-1-1")[0] L_previous = None for i in range(10000): np.random.seed(0) mx.random.seed(0) L = nd.linalg.potrf(K) if L_previous is not None: # print("Result difference, should be 0:", nd.sum(nd.abs(L-L_previous))) is_unequal = nd.sum(1 - nd.equal(L, L_previous)) # 0 if same if is_unequal: print ('ERROR') break L_previous = L ```
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
