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 14, 05:30 UTC
[V0 Deprecation] Remove unused output processor util (#25023)
Signed-off-by: Woosuk Kwon <woosuk@thinkingmachines.ai>
Woosuk Kwon committed
Sep 17, 2025 at 09:50 UTC
99cc41ad50c08e745571abe568226f9fcae61ccd
1 file changed
-28
vllm/engine/output_processor/util.py
deleted
-28
@@ -1,28 +0,0 @@
1
-# SPDX-License-Identifier: Apache-2.0
2
-# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
3
-
4
-from typing import List
5
-from typing import Sequence as GenericSequence
6
-from typing import cast
7
-
8
-from vllm.model_executor.layers.sampler import SamplerOutput
9
-from vllm.sequence import CompletionSequenceGroupOutput, SequenceGroupOutput
10
-
11
-
12
-def create_output_by_sequence_group(
13
- outputs: GenericSequence[SamplerOutput],
14
- num_seq_groups: int) -> List[List[SequenceGroupOutput]]:
15
- """Helper method which transforms a 2d list organized by
16
- [step][sequence group] into [sequence group][step].
17
- """
18
- output_by_sequence_group: List[List[CompletionSequenceGroupOutput]] = [
19
- [] for _ in range(num_seq_groups)
20
- ]
21
- for step in outputs:
22
- sequence_group_output: CompletionSequenceGroupOutput
23
- for i, sequence_group_output in enumerate(step):
24
- output_by_sequence_group[i].append(sequence_group_output)
25
-
26
- # Cast to the more generic type that CompletionSequenceGroupOutput
27
- # inherits from.
28
- return cast(List[List[SequenceGroupOutput]], output_by_sequence_group)