Skip to content

[test] recovered activation checkpointig test #1459

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
FrankLeeeee wants to merge 3 commits into from
Closed

[test] recovered activation checkpointig test #1459

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
06235d2
[test] recovered activation checkpointig test
FrankLeeeee Aug 16, 2022
0b82ac3
polish code
FrankLeeeee Aug 16, 2022
367ff8e
polish code
FrankLeeeee Aug 16, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
10 changes: 8 additions & 2 deletions tests/test_utils/test_activation_checkpointing.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,7 +6,7 @@
import torch.nn.functional as F
from colossalai.context.parallel_mode import ParallelMode
from colossalai.context.random import add_seed, seed, set_mode, reset_seeds
from colossalai.utils import checkpoint
from colossalai.utils.activation_checkpoint import checkpoint


def forward(x, weight):
Expand All @@ -17,9 +17,10 @@ def forward(x, weight):


@pytest.mark.gpu
@pytest.mark.skip("set seed error")
@pytest.mark.parametrize("cpu_offload", [True, False])
def test_activation_checkpointing(cpu_offload):
# clear all previous seeds possibly set by other tests
reset_seeds()

# We put initilization here to avoid change cuda rng state below
inputs = torch.rand(2, 2, requires_grad=True, device='cuda')
Expand Down Expand Up @@ -62,3 +63,8 @@ def test_activation_checkpointing(cpu_offload):
# other tests will fail if running together with this test
# as other tests can't overwrite the seed set by this test
reset_seeds()


if __name__ == '__main__':
test_activation_checkpointing(cpu_offload=False)
test_activation_checkpointing(cpu_offload=True)