A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 15, 06:00 UTC
[Fix] The zip function in Python 3.9 does not have the strict argument (#19549)
Signed-off-by: 汪志鹏 <wangzhipeng628@gmail.com>
汪志鹏 committed
Jun 13, 2025 at 14:57 UTC
cefdb9962d788393f96f8881e0e3c1434ac09c2c
2 files changed
+4
-8
vllm/v1/worker/gpu_model_runner.py
+2
-4
@@ -467,10 +467,8 @@ class GPUModelRunner(LoRAModelRunnerMixin):
467
# Update the block IDs.
468
if not req_data.resumed_from_preemption:
469
# Append the new blocks to the existing block IDs.
470
- for block_ids, new_block_ids in zip( # type: ignore[call-overload]
471
- req_state.block_ids,
472
- req_data.new_block_ids,
473
- strict=True):
470
+ for block_ids, new_block_ids in zip(req_state.block_ids,
471
+ req_data.new_block_ids):
472
block_ids.extend(new_block_ids)
473
else:
474
# The request is resumed from preemption.
vllm/v1/worker/tpu_model_runner.py
+2
-4
@@ -413,10 +413,8 @@ class TPUModelRunner(LoRAModelRunnerMixin):
413
req_state.num_computed_tokens = req_data.num_computed_tokens
414
if not req_data.resumed_from_preemption:
415
# Append the new blocks to the existing block IDs.
416
- for block_ids, new_block_ids in zip( # type: ignore[call-overload]
417
- req_state.block_ids,
418
- req_data.new_block_ids,
419
- strict=True):
416
+ for block_ids, new_block_ids in zip(req_state.block_ids,
417
+ req_data.new_block_ids):
418
block_ids.extend(new_block_ids)
419
else:
420
# The request is resumed from preemption.