connorgoggins commented on a change in pull request #17675: [Large Tensor] Fix
multi_lars op
URL: https://github.com/apache/incubator-mxnet/pull/17675#discussion_r383560463
##########
File path: tests/nightly/test_large_array.py
##########
@@ -455,6 +455,20 @@ def npy_instance_norm(data, gamma, beta, axis, eps=1E-5):
assert_almost_equal(out, out_nd.asnumpy(), forward_check_eps,
forward_check_eps)
+ def check_multi_lars():
+ lrs = nd.random_normal(shape=(2**32 + 1, 1))
+ weights_sum_sq = nd.random_normal(shape=(2**32 + 1, 1))
+ grads_sum_sq = nd.random_normal(shape=(2**32 + 1, 1))
+ wds = nd.random_normal(shape=(2**32 + 1, 1))
+ eta = .1
+ eps = .9
+
+ out = nd.multi_lars(lrs=lrs, weights_sum_sq=weights_sum_sq,
grads_sum_sq=grads_sum_sq,
+ wds=wds, eta=eta, eps=eps)
+
+ assert out.shape[0] == 4294967297
Review comment:
Done!
----------------------------------------------------------------
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:
us...@infra.apache.org
With regards,
Apache Git Services
