ci: unify PR test suite naming (#21187)
This commit is contained in:
@@ -21,7 +21,7 @@ from sglang.test.test_utils import (
|
||||
)
|
||||
|
||||
# FlashMLA attention backend tests with MTP speculative decoding
|
||||
register_cuda_ci(est_time=284, suite="stage-b-test-large-1-gpu")
|
||||
register_cuda_ci(est_time=284, suite="stage-b-test-1-gpu-large")
|
||||
|
||||
|
||||
class TestFlashMLAAttnBackend(unittest.TestCase):
|
||||
|
||||
@@ -13,8 +13,8 @@ from sglang.test.test_utils import (
|
||||
)
|
||||
|
||||
# MLA attention test with MGSM evaluation
|
||||
register_cuda_ci(est_time=194, suite="stage-b-test-large-1-gpu")
|
||||
register_amd_ci(est_time=1100, suite="stage-b-test-small-1-gpu-amd")
|
||||
register_cuda_ci(est_time=194, suite="stage-b-test-1-gpu-large")
|
||||
register_amd_ci(est_time=1100, suite="stage-b-test-1-gpu-small-amd")
|
||||
|
||||
|
||||
class TestMLA(CustomTestCase):
|
||||
|
||||
@@ -16,10 +16,10 @@ from sglang.test.test_utils import (
|
||||
)
|
||||
|
||||
# DeepSeek-V3 MLA tests with torch compile, FA3, and MTP speculative decoding
|
||||
register_cuda_ci(est_time=442, suite="stage-b-test-large-1-gpu")
|
||||
register_cuda_ci(est_time=442, suite="stage-b-test-1-gpu-large")
|
||||
register_amd_ci(
|
||||
est_time=221,
|
||||
suite="stage-b-test-small-1-gpu-amd",
|
||||
suite="stage-b-test-1-gpu-small-amd",
|
||||
disabled="see https://github.com/sgl-project/sglang/issues/12574",
|
||||
)
|
||||
|
||||
|
||||
@@ -15,7 +15,7 @@ from sglang.test.test_utils import (
|
||||
)
|
||||
|
||||
# FlashInfer MLA backend tests with MTP speculative decoding
|
||||
register_cuda_ci(est_time=302, suite="stage-b-test-large-1-gpu")
|
||||
register_cuda_ci(est_time=302, suite="stage-b-test-1-gpu-large")
|
||||
|
||||
|
||||
class TestFlashinferMLA(CustomTestCase):
|
||||
|
||||
@@ -13,8 +13,8 @@ from sglang.test.test_utils import (
|
||||
)
|
||||
|
||||
# MLA FP8 KV cache test with MGSM evaluation
|
||||
register_cuda_ci(est_time=77, suite="stage-b-test-large-1-gpu")
|
||||
register_amd_ci(est_time=800, suite="stage-b-test-small-1-gpu-amd")
|
||||
register_cuda_ci(est_time=77, suite="stage-b-test-1-gpu-large")
|
||||
register_amd_ci(est_time=800, suite="stage-b-test-1-gpu-small-amd")
|
||||
|
||||
|
||||
class TestMLA(CustomTestCase):
|
||||
|
||||
@@ -16,7 +16,7 @@ from sglang.test.test_utils import (
|
||||
)
|
||||
|
||||
# DeepSeek-V3 INT8 quantization tests (channel and block INT8)
|
||||
register_cuda_ci(est_time=341, suite="stage-b-test-large-1-gpu")
|
||||
register_cuda_ci(est_time=341, suite="stage-b-test-1-gpu-large")
|
||||
|
||||
|
||||
class TestMLADeepseekV3ChannelInt8(CustomTestCase):
|
||||
|
||||
Reference in New Issue
Block a user