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

[CI] Prune back the number of tests in tests/kernels/* (#9932)

Browse Source 
Signed-off-by: mgoin <michael@neuralmagic.com>
This commit is contained in:
Michael Goin
2024-11-05 16:02:32 -05:00
committed by GitHub
parent 02462465ea
commit 235366fe2e
10 changed files with 60 additions and 36 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
tests/kernels/test_blocksparse_attention.py
View File

@@ -25,10 +25,10 @@ PARTITION_SIZE = 512
DTYPES = [torch.half, torch.bfloat16]
NUM_GEN_SEQS = [3] # Arbitrary values for testing
NUM_PREFILL_SEQS = [3] # Arbitrary values for testing
NUM_HEADS = [(40, 40), (64, 8)] # Arbitrary values for testing
NUM_HEADS = [(40, 40)] # Arbitrary values for testing
HEAD_SIZES = [64, 112]
BLOCK_SIZES = [16, 32]
BLOCK_SIZES = [16]
USE_ALIBI = [False, True]
KV_CACHE_DTYPE = ["auto", "fp8"]
SEEDS = [0]
@@ -37,7 +37,7 @@ BLOCKSPARSE_LOCAL_BLOCKS = [16]
BLOCKSPARSE_VERT_STRIDES = [8]
BLOCKSPARSE_BLOCK_SIZES = [64]
BLOCKSPARSE_HEADS_SLIDINGS = [0, 2, -1]
BLOCKSPARSE_HEADS_SLIDINGS = [2, -1]
BLOCKSPARSE_HOMO_HEADS = [True, False]