github/ColossalAI
1
0
Fork 0
mirror of https://github.com/hpcaitech/ColossalAI.git synced 2025-09-05 11:02:05 +00:00
Code Issues Packages Projects Releases Wiki Activity

[test] fixed rerun_on_exception and adapted test cases (#487)

Browse Source
This commit is contained in:
Frank Lee
2022-03-25 17:25:12 +08:00
committed by GitHub
parent 4d322b79da
commit 3601b2bad0
31 changed files with 143 additions and 135 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
tests/test_moe/test_kernel.py
View File

@@ -9,6 +9,7 @@ from colossalai.core import global_context as gpc
from colossalai.utils import free_port, get_current_device
from colossalai.nn.layer.moe import Top1Router, Top2Router, MoeLayer, Experts
from colossalai.context.moe_context import MOE_CONTEXT
from colossalai.testing import rerun_on_exception
BATCH_SIZE = 16
NUM_EXPERTS = 4
@@ -86,6 +87,7 @@ def run_routing(rank, world_size, port, rs=2, hidden_size=128, data_type=torch.f
@pytest.mark.parametrize("hidden_size", [32, 144])
@pytest.mark.parametrize("data_type", [torch.float32, torch.float16])
@pytest.mark.parametrize("router", [Top1Router, Top2Router])
@rerun_on_exception(exception_type=mp.ProcessRaisedException, pattern=".*Address already in use.*")
def test_moe_kernel(rs, hidden_size, data_type, router):
world_size = 4
run_func = partial(run_routing,