[Bugfix][KV Offload] count appended GPU blocks in store group_sizes (#42945)

Signed-off-by: Kfir Toledo <kfir.toledo@ibm.com>
This commit is contained in:
Kfir Toledo
2026-05-18 11:36:02 +00:00
committed by GitHub
parent df852ed503
commit e414e1f1c0
@@ -710,7 +710,6 @@ class OffloadingConnectorScheduler:
offloaded_block_idx = start_block_idx + idx
gpu_block_idx = offloaded_block_idx * block_size_factor
num_group_blocks += block_size_factor
for i in range(block_size_factor):
block_id = block_ids[gpu_block_idx + i]
if block_id == 0:
@@ -720,6 +719,7 @@ class OffloadingConnectorScheduler:
elif start_gpu_block_idx is None:
start_gpu_block_idx = gpu_block_idx + i
src_block_ids.append(block_id)
num_group_blocks += 1
if is_sliding_window:
sliding_window_block_ids.append(block_id)
else: