[Bugfix][Core] Reserve the KV null block when validating max_model_len (#47272)
Signed-off-by: hyungjun oh <92hyungjun@gmail.com> Signed-off-by: Nick Hill <nickhill123@gmail.com> Co-authored-by: Claude <noreply@anthropic.com> Co-authored-by: Nick Hill <nickhill123@gmail.com>
This commit is contained in:
co-authored by
Claude
Nick Hill
parent
c0233fcf01
commit
76fb6d210a
@@ -1244,7 +1244,7 @@ def test_prompt_logprobs_with_chunking_and_preemption():
|
||||
max_model_len=512,
|
||||
enable_chunked_prefill=True,
|
||||
max_num_batched_tokens=48, # Force prefill chunking
|
||||
num_gpu_blocks_override=32, # Force preemptions
|
||||
num_gpu_blocks_override=33, # Force preemptions (32 usable + null block)
|
||||
disable_log_stats=False,
|
||||
gpu_memory_utilization=0.25,
|
||||
) as vllm_model:
|
||||
|
||||
Reference in New Issue
Block a user