mirror of
https://git.datalinker.icu/vllm-project/vllm.git
synced 2026-01-19 14:44:28 +08:00
e489ad7a21
- **Add SPDX license headers to python source files**
- **Check for SPDX headers using pre-commit**
commit 9d7ef44c3cfb72ca4c32e1c677d99259d10d4745
Author: Russell Bryant <rbryant@redhat.com>
Date: Fri Jan 31 14:18:24 2025 -0500
Add SPDX license headers to python source files
This commit adds SPDX license headers to python source files as
recommended to
the project by the Linux Foundation. These headers provide a concise way
that is
both human and machine readable for communicating license information
for each
source file. It helps avoid any ambiguity about the license of the code
and can
also be easily used by tools to help manage license compliance.
The Linux Foundation runs license scans against the codebase to help
ensure
we are in compliance with the licenses of the code we use, including
dependencies. Having these headers in place helps that tool do its job.
More information can be found on the SPDX site:
- https://spdx.dev/learn/handling-license-info/
Signed-off-by: Russell Bryant <rbryant@redhat.com>
commit 5a1cf1cb3b80759131c73f6a9dddebccac039dea
Author: Russell Bryant <rbryant@redhat.com>
Date: Fri Jan 31 14:36:32 2025 -0500
Check for SPDX headers using pre-commit
Signed-off-by: Russell Bryant <rbryant@redhat.com>
---------
Signed-off-by: Russell Bryant <rbryant@redhat.com>
45 lines
2.0 KiB
Python
45 lines
2.0 KiB
Python
# SPDX-License-Identifier: Apache-2.0
|
|
|
|
from typing import List, Optional
|
|
|
|
from vllm.sequence import SequenceGroupMetadata
|
|
from vllm.worker.model_runner_base import (ModelRunnerBase,
|
|
ModelRunnerInputBase,
|
|
ModelRunnerWrapperBase)
|
|
|
|
|
|
class TargetModelRunner(ModelRunnerWrapperBase):
|
|
"""Specialized model runner for speculative decoding target model.
|
|
In speculative decoding, the log probabilities selected finally may not
|
|
be the same ones as selected by the target model sampling. This means
|
|
that the time spent in the log probability calculation of the target model
|
|
is time wasted, since we calculate log probabilities after deciding which
|
|
tokens are accepted. For this reason disabling log probabilities in the
|
|
target model will make decode faster. The model runner sets the
|
|
SamplingMetadata parameters according to whether log probabilities are
|
|
requested or not.
|
|
"""
|
|
|
|
def __init__(self, model_runner: ModelRunnerBase):
|
|
# An internal boolean member variable to indicate if token log
|
|
# probabilities are needed or not.
|
|
super().__init__(model_runner)
|
|
self.disable_logprobs = True
|
|
|
|
def prepare_model_input(
|
|
self,
|
|
seq_group_metadata_list: List[SequenceGroupMetadata],
|
|
virtual_engine: int = 0,
|
|
finished_requests_ids: Optional[List[str]] = None,
|
|
) -> ModelRunnerInputBase:
|
|
model_input: ModelRunnerInputBase =\
|
|
self.model_runner.prepare_model_input(
|
|
seq_group_metadata_list, virtual_engine, finished_requests_ids)
|
|
# If token log probabilities is disabled then skip generating sampler
|
|
# CPU output. We directly serialize the GPU sampled_token_id tensors
|
|
# as needed. If log probabilities is enabled then synchronize all the
|
|
# sampling related tensors which includes the logprobs tensors.
|
|
model_input.sampling_metadata.skip_sampler_cpu_output = (
|
|
self.disable_logprobs)
|
|
return model_input
|