biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

kernels/moe test pruning (#27053)

Browse Source 
Signed-off-by: Fardin Hoque <kfhfar@amazon.com>
Co-authored-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>
This commit is contained in:
Fardin Hoque
2025-10-29 21:10:34 -07:00
committed by GitHub
parent 17d055f527
commit b8c48c5d72
13 changed files with 34 additions and 56 deletions
Download Patch File Download Diff File Expand all files Collapse all files

14
tests/kernels/moe/test_block_fp8.py
View File

@@ -42,57 +42,43 @@ DTYPES = [torch.bfloat16] # [torch.half, torch.bfloat16, torch.float32]
# and its hidden size is 7168.
MNK_FACTORS = [
(1, 128, 128),
(1, 512, 512),
(1, 128, 7168),
(1, 1024, 7168),
(1, 4608, 128),
(1, 4608, 512),
(1, 4608, 7168),
(83, 128, 128),
(83, 512, 512),
(83, 1024, 7168),
(83, 4608, 512),
(83, 4608, 7168),
(128, 128, 128),
(128, 512, 512),
(128, 1024, 7168),
(128, 4608, 512),
(128, 4608, 7168),
(2048, 128, 128),
(2048, 1024, 7168),
(2048, 4608, 512),
(2048, 4608, 7168),
(8192, 128, 128),
(8192, 512, 512),
(8192, 128, 7168),
(8192, 1024, 7168),
(8192, 4608, 512),
(8192, 4608, 7168),
]
MNK_FACTORS_DG = [
(128, 128, 128),
(128, 512, 512),
(128, 128, 7168),
(128, 1024, 7168),
(128, 4608, 128),
(128, 4608, 512),
(128, 4608, 7168),
(192, 128, 128),
(192, 512, 512),
(192, 1024, 7168),
(192, 4608, 512),
(192, 4608, 7168),
(1335, 128, 128),
(1335, 1024, 7168),
(1335, 4608, 512),
(1335, 4608, 7168),
(2048, 128, 128),
(2048, 512, 512),
(2048, 128, 7168),
(2048, 1024, 7168),
(2048, 4608, 128),
(2048, 4608, 512),
(2048, 4608, 7168),
]