github/ColossalAI
1
0
Fork 0
mirror of https://github.com/hpcaitech/ColossalAI.git synced 2025-09-16 06:30:41 +00:00
Code Issues Packages Projects Releases Wiki Activity

[hotfix] skipped unsafe test cases (#1282)

Browse Source
This commit is contained in:
Frank Lee
2022-07-13 00:08:59 +08:00
committed by GitHub
parent 79fe7b027a
commit 7e8114a8dd
14 changed files with 22 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
tests/test_utils/test_checkpoint/test_checkpoint_1d.py
View File

@@ -67,6 +67,7 @@ def check_checkpoint_1d(rank, world_size, port):
@pytest.mark.dist
@pytest.mark.skip("takes too long")
@skip_if_not_enough_gpus(min_gpus=8)
@rerun_on_exception(exception_type=mp.ProcessRaisedException, pattern=".*Address already in use.*")
def test_checkpoint_1d():

1
tests/test_utils/test_checkpoint/test_checkpoint_2d.py
View File

@@ -67,6 +67,7 @@ def check_checkpoint_2d(rank, world_size, port):
@pytest.mark.dist
@pytest.mark.skip("takes too long")
@skip_if_not_enough_gpus(min_gpus=8)
@rerun_on_exception(exception_type=mp.ProcessRaisedException, pattern=".*Address already in use.*")
def test_checkpoint_2d():

2
tests/test_utils/test_checkpoint/test_checkpoint_2p5d.py
View File

@@ -37,6 +37,7 @@ def build_pipeline(model):
def check_equal(A, B):
assert torch.allclose(A, B, rtol=1e-3, atol=1e-2)
def check_checkpoint_2p5d(rank, world_size, port):
config = dict(parallel=dict(pipeline=dict(size=2), tensor=dict(size=4, depth=1, mode="2.5d")),)
@@ -66,6 +67,7 @@ def check_checkpoint_2p5d(rank, world_size, port):
@pytest.mark.dist
@pytest.mark.skip("takes too long")
@skip_if_not_enough_gpus(min_gpus=8)
@rerun_on_exception(exception_type=mp.ProcessRaisedException, pattern=".*Address already in use.*")
def test_checkpoint_2p5d():

2
tests/test_utils/test_checkpoint/test_checkpoint_3d.py
View File

@@ -37,6 +37,7 @@ def build_pipeline(model):
def check_equal(A, B):
assert torch.allclose(A, B, rtol=1e-3, atol=1e-2)
def check_checkpoint_3d(rank, world_size, port):
config = dict(parallel=dict(pipeline=dict(size=1), tensor=dict(size=8, mode="3d")),)
@@ -66,6 +67,7 @@ def check_checkpoint_3d(rank, world_size, port):
@pytest.mark.dist
@pytest.mark.skip("takes too long")
@skip_if_not_enough_gpus(min_gpus=8)
@rerun_on_exception(exception_type=mp.ProcessRaisedException, pattern=".*Address already in use.*")
def test_checkpoint_3d():