[Attention] Cache attention metadata builds across hybrid KV-cache groups (#29627)

Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
Co-authored-by: Stanislaw Wozniak <stw@zurich.ibm.com>
This commit is contained in:
Lucas Wilkinson
2025-12-16 17:10:16 -05:00
committed by GitHub
co-authored by Stanislaw Wozniak
parent 254a7f8fd6
commit 9fec0e13d5
7 changed files with 105 additions and 13 deletions
@@ -172,7 +172,7 @@ def test_local_attention_virtual_batches(test_data: LocalAttentionTestData):
)
# Call the function
result = make_local_attention_virtual_batches(
result, _ = make_local_attention_virtual_batches(
attn_chunk_size, common_attn_metadata, block_size
)