Summary:
4GB is still too large and leads to CUDA OOM failures.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/15959
Differential Revision:
D13635146
Pulled By: mrshenli
fbshipit-source-id:
3dc34a03d6ed65c458839d8fa37cd05bf3bc8106
]
tri_large_tests_args = [
- (1, 268435455),
# Large test cases below are deliberately commented out to speed up CI
# tests and to avoid OOM error. When modifying implementations of
# tril_indices and triu_indices, please enable these tests and make sure
# they pass.
#
+ # (1, 268435455),
# (5000, 5000),
# (10000, 10000),
# (268435455, 1),