c980e68d40
### What this PR does / why we need it?
This PR aims to support aclgraph for model runner v2, please see RFC
#5208. The PR contains these modifications:
- adapt to newest commit of vllm main branch.
- supply a unified interface of extra forward context for both model
runner v1 and model runner v2.
- implement graph mode for main model.
### Does this PR introduce _any_ user-facing change?
no
### How was this patch tested?
- vLLM version: v0.16.0
- vLLM main:
4034c3d32e
---------
Signed-off-by: Ronald1995 <ronaldautomobile@163.com>
81 lines
2.8 KiB
Python
81 lines
2.8 KiB
Python
# Adapt from https://github.com/vllm-project/vllm/blob/main/vllm/v1/worker/gpu/sample/sampler.py.
|
|
# Copyright (c) 2025 Huawei Technologies Co., Ltd. All Rights Reserved.
|
|
#
|
|
# 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.
|
|
#
|
|
import numpy as np
|
|
import torch
|
|
from vllm.v1.worker.gpu.sample.sampler import Sampler
|
|
|
|
from vllm_ascend.worker.v2.sample.gumbel import gumbel_sample
|
|
|
|
|
|
class AscendSampler(Sampler):
|
|
def sample(
|
|
self,
|
|
logits: torch.Tensor,
|
|
idx_mapping: torch.Tensor,
|
|
idx_mapping_np: np.ndarray,
|
|
pos: torch.Tensor,
|
|
input_ids: torch.Tensor,
|
|
expanded_local_pos: torch.Tensor,
|
|
) -> tuple[torch.Tensor, torch.Tensor]:
|
|
"""Override sample method because we need to override triton operators
|
|
called in the method.
|
|
"""
|
|
# Copy logits to a new FP32 tensor.
|
|
logits = torch.empty_like(logits, dtype=torch.float32).copy_(logits)
|
|
|
|
# Apply logit bias (e.g., allowed_token_ids, min_tokens) in place.
|
|
self.logit_bias_state.apply_logit_bias(logits, idx_mapping, idx_mapping_np, pos)
|
|
|
|
# Apply penalties in place.
|
|
self.penalties_state.apply_penalties(
|
|
logits,
|
|
idx_mapping,
|
|
idx_mapping_np,
|
|
input_ids,
|
|
expanded_local_pos,
|
|
self.num_speculative_tokens,
|
|
)
|
|
|
|
# Apply bad words masking in place.
|
|
self.bad_words_state.apply_bad_words(
|
|
logits,
|
|
idx_mapping,
|
|
idx_mapping_np,
|
|
input_ids,
|
|
expanded_local_pos,
|
|
)
|
|
|
|
# Apply temperature in place.
|
|
self.sampling_states.apply_temperature(logits, idx_mapping, idx_mapping_np)
|
|
|
|
# Apply min_p in place.
|
|
self.sampling_states.apply_min_p(logits, idx_mapping, idx_mapping_np)
|
|
|
|
# Apply top_k and/or top_p. This might or might not return a new tensor.
|
|
logits = self.sampling_states.apply_top_k_top_p(logits, idx_mapping, idx_mapping_np)
|
|
|
|
# Sample the next token.
|
|
sampled = gumbel_sample(
|
|
logits,
|
|
idx_mapping,
|
|
self.sampling_states.temperature.gpu,
|
|
self.sampling_states.seeds.gpu,
|
|
pos,
|
|
apply_temperature=False,
|
|
)
|
|
return sampled, logits
|