[V1] reduce block size for tree attention correctness test to fix 'ou… (#22207)

Signed-off-by: Giancarlo Delfin <gdelfin@meta.com>
2025-08-04 19:11:06 -07:00
parent 7175817637
commit 5ea71ff46f
1 changed files with 1 additions and 1 deletions
--- a/tests/v1/spec_decode/test_tree_attention.py
+++ b/tests/v1/spec_decode/test_tree_attention.py
@ -155,7 +155,7 @@ def test_tree_attn_correctness() -> None:

    dim_per_head = 128
    num_kv_heads = 2
-    block_size = 128
+    block_size = 32
    max_sequence_length = 8192
    randomize_blocks = True
    for batch_size in [1, 16, 32]: