This pull request refactors the speculative decoding proposer interface
to align with upstream vLLM, removing the local `Proposer` interface and
renaming methods to `propose`.
This is the first step. In the future we should remove the class
register and just add few Ascend specified method once the arch in vLLM
is ready.
- vLLM version: v0.16.0
- vLLM main:
15d76f74e2
Signed-off-by: wangxiyuan <wangxiyuan1007@gmail.com>
39 lines
1.7 KiB
Python
39 lines
1.7 KiB
Python
#
|
|
# Copyright (c) 2025 Huawei Technologies Co., Ltd. All Rights Reserved.
|
|
# Copyright 2023 The vLLM team.
|
|
#
|
|
# Licensed under the Apache License, Version 2.0 (the "License");
|
|
# you may not use this file except in compliance with the License.
|
|
# You may obtain a copy of the License at
|
|
#
|
|
# http://www.apache.org/licenses/LICENSE-2.0
|
|
#
|
|
# Unless required by applicable law or agreed to in writing, software
|
|
# distributed under the License is distributed on an "AS IS" BASIS,
|
|
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
# See the License for the specific language governing permissions and
|
|
# limitations under the License.
|
|
# This file is a part of the vllm-ascend project.
|
|
# Adapted from vllm-project/vllm/vllm/worker/gpu_model_runner.py
|
|
#
|
|
from vllm_ascend.spec_decode.eagle_proposer import AscendEagleProposer
|
|
from vllm_ascend.spec_decode.medusa_proposer import AscendMedusaProposer
|
|
from vllm_ascend.spec_decode.mtp_proposer import AscendMtpProposer
|
|
from vllm_ascend.spec_decode.ngram_proposer import AscendNgramProposer
|
|
from vllm_ascend.spec_decode.suffix_proposer import AscendSuffixDecodingProposer
|
|
|
|
|
|
def get_spec_decode_method(method, vllm_config, device, runner):
|
|
if method == "ngram":
|
|
return AscendNgramProposer(vllm_config, runner)
|
|
elif method == "suffix":
|
|
return AscendSuffixDecodingProposer(vllm_config, runner)
|
|
elif method == "medusa":
|
|
return AscendMedusaProposer(vllm_config, device)
|
|
elif method in ("eagle", "eagle3"):
|
|
return AscendEagleProposer(vllm_config, device, runner)
|
|
elif method == "mtp":
|
|
return AscendMtpProposer(vllm_config, device, runner)
|
|
else:
|
|
raise ValueError(f"Unknown speculative decoding method: {method}")
|