54c0e63df7
<!-- Thanks for sending a pull request! BEFORE SUBMITTING, PLEASE READ https://docs.vllm.ai/en/latest/contributing/overview.html --> ### What this PR does / why we need it? As custom deepseek modeling do some changes to support graph mode in https://github.com/vllm-project/vllm-ascend/pull/585, so i follow it to change custom deepseek_mtp modeling. And some modifications for k>1 were not carried over by the https://github.com/vllm-project/vllm-ascend/pull/429, now i add it. In order to better take care of the MTP feature in the vllm-ascend repository, I added cases related to graph mode(torchair), but i skip it since torchair can not correctly clean up memory in vllmrunner. Also i add some case for MTP quantization weights, but test weight is not ready, so i skip it and i will open it when test quant weights is ready. https://github.com/vllm-project/vllm-ascend/pull/648 did not completely fix the sample change(https://github.com/vllm-project/vllm-ascend/issues/660) issue, I added the relevant changes. ### Does this PR introduce _any_ user-facing change? now, u can use following method to use mtp in deepseek v3/r1 float or quant weights with eager mode. ```python llm = LLM( model="wemaster/deepseek_mtp_main_random_bf16", tensor_parallel_size=2, speculative_config={ "num_speculative_tokens": 1, }, enforce_eager=True, trust_remote_code=True, disable_log_stats=False, gpu_memory_utilization=0.8, max_model_len=64, ) ``` or use mtp in deepseek v3/r1 float or quant weights with graph mode（torchair） ```python llm = LLM( model="wemaster/deepseek_mtp_main_random_bf16", tensor_parallel_size=2, speculative_config={ "num_speculative_tokens": 1, }, trust_remote_code=True, additional_config={ 'enable_graph_mode': True, }, disable_log_stats=False, gpu_memory_utilization=0.8, max_model_len=64, ) ``` add notes: 1. now, we support k>1, so u can set num_speculative_tokens > 1 if there is sufficient redundant computing power; 2. MTP is not supported in V1, we will support it when vLLM does it in https://github.com/vllm-project/vllm/issues/13500. 3. if u run MTP failed by `segmentation fault`, u can follow v0.7.3 patch https://github.com/vllm-project/vllm-ascend/pull/236 file `vllm_ascend/patch/patch_metrics.py` method `__npu_async_metrics_collector_init__` ### How was this patch tested? local tested passed and test by CI Signed-off-by: mengwei805 <mengwei25@huawei.com>
212 lines
11 KiB
Python
212 lines
11 KiB
Python
#
|
||
# Copyright (c) 2025 Huawei Technologies Co., Ltd. All Rights Reserved.
|
||
# This file is a part of the vllm-ascend project.
|
||
#
|
||
# 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 module manage the patch for vllm. There are two folders in this module:
|
||
# - platform: contains the patches applied before worker starts. It's called by
|
||
# `vllm_ascend.utils.adapt_patch(is_global_patch=True)` in
|
||
# `vllm_ascend.platform.NPUPlatform.pre_register_and_update()` function.
|
||
# - worker: contains the patches applied when worker starts. It's called by
|
||
# `vllm_ascend.utils.adapt_patch(is_global_patch=False)` in
|
||
# each worker's `__init__` function.
|
||
#
|
||
# Then in each kind of patch, there are three folders:
|
||
# - patch_0_8_4: contains the patches applied when vllm version is 0.8.4.
|
||
# - patch_main: contains the patches applied when vllm version is main branch.
|
||
# - patch_common: contains the patches applied in both 0.8.4 and main branch.
|
||
#
|
||
# In the future, with the vllm version upgrade, the new patch folder such as
|
||
# patch_0_8_5, patch_0_8_6, etc. will be added to manage the patch for different
|
||
# vllm version. And the patch_common will contain the patches applied in all the
|
||
# vllm version.
|
||
# Once the vllm version is too old that vllm-ascend will not support, the related
|
||
# patch folder will be removed as well.
|
||
#
|
||
# Once a new patch is added in vllm-ascend, please add the patch description into this file as well.
|
||
# ----------------------------------------------------------------------------------
|
||
|
||
# What's Patched and how it works:
|
||
# --------------------------------
|
||
# * Platform Patch:
|
||
# =================
|
||
# ** File: platform/patch_0_8_4/patch_config.py**
|
||
# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
|
||
# 1. `vllm.config.ModelConfig.__init__()`
|
||
# Why:
|
||
# It is hard coded for sleep mode to support cuda platform only
|
||
# How：
|
||
# Using a new method to check if sleep mode is available
|
||
# Related PR (if no, explain why): 1. refused by vllm. 2. vllm doesn't support 3. prepare to submit....
|
||
# https://github.com/vllm-project/vllm/pull/16562
|
||
# Future Plan:
|
||
# This patch is only used for 084 and can't be revert. just keep as it is.
|
||
#
|
||
# ** File: platform/patch_common/patch_distributed.py**
|
||
# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
|
||
# 1. `vllm.distributed.parallel_state.destroy_model_parallel()`
|
||
# Why:
|
||
# vllm dose not support outside platform maintain its own `CoordinatorGroup`, vllm-ascend maintain EP and ETP
|
||
# inside of the repo, and needs a common interface to destroy them, this patch add the interface of destroy
|
||
# platform owned `CoordinatorGroup` to make sure all the CoordinateGroup can be properly destroyed
|
||
# How：
|
||
# Call platform method `destroy_platform_model_parallel` to destroy all the `CoordinateGroup`
|
||
# Related PR (if no, explain why): no related PR, we want add this ability into vllm
|
||
# Future Plan:
|
||
# Remove those patch when vllm merged them
|
||
# 2. `vllm.distributed.stateless_init_torch_distributed_process_group()`
|
||
# Why:
|
||
# The stateless process group can not be initialized except from gloo and nccl backend, vllm-ascend
|
||
# needs to initialize its own stateless process group for communication, so we add the platform related
|
||
# call to the `stateless_init_torch_distributed_process_group`, to enable other platform which may support
|
||
# stateless process group initialize method
|
||
# How：
|
||
# Call platform method `platform_has_backend_register` to judge if there is a stateless process group initialize
|
||
# method and call platform method `platform_register_backend` to initialize them
|
||
# Related PR (if no, explain why): no related PR, we want add this ability into vllm
|
||
# Future Plan:
|
||
# Remove those patch when vllm merged them
|
||
# 3. `ParallelConfig.get_next_dp_init_port`
|
||
# Why:
|
||
# We want to get dp port from env variable, so the multi-node inference can be properly initialized and run.
|
||
# How：
|
||
# Get the dp port from env variable enable multi-mode dp inference
|
||
# Related PR (if no, explain why): no related PR, we want add this ability into vllm
|
||
# Future Plan:
|
||
# Its a workaround in vllm-ascend to enable multi-node dp inference, maybe removed if vllm have better plan
|
||
# on multi-node dp inference implementation
|
||
# 4. `ParallelConfig.stateless_init_dp_group`
|
||
# Why:
|
||
# vLLM use gloo backend by default to initialize stateless dp process gourp, but we want to use hccl here to
|
||
# get better performance
|
||
# How：
|
||
# adopt nccl backend to init process group
|
||
# Related PR (if no, explain why): no related PR, we want add this ability into vllm
|
||
# Future Plan:
|
||
# Remove those patch when vllm merged them
|
||
#
|
||
#
|
||
# * Worker Patch:
|
||
# ===============
|
||
# ** File: worker/patch_0_8_4/patch_metrics.py **
|
||
# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
|
||
# 1. `vllm.spec_decode.metrics.AsyncMetricsCollector.init_tensors` and
|
||
# `vllm.spec_decode.metrics.AsyncMetricsCollector._copy_rejsample_metrics_async`
|
||
# Why:
|
||
# There are cuda hard code (torch.cuda.Stream) in `AsyncMetricsCollector.init_tensors` and
|
||
# `AsyncMetricsCollector._copy_rejsample_metrics_async`
|
||
# How：
|
||
# Replace it with the corresponding npu method
|
||
# Related PR (if no, explain why): 1. refused by vllm. 2. vllm doesn't support 3. prepare to submit....
|
||
# https://github.com/vllm-project/vllm/pull/14411
|
||
# Future Plan:
|
||
# Revert it when the related pr is merged in vllm.
|
||
#
|
||
# ** File: worker/patch_0_8_4/patch_spec_decode_worker.py **
|
||
# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
|
||
# 1. `vllm.spec_decode.spec_decode_worker.SpecDecodeWorker._configure_model_sampler_for_spec_decode`
|
||
# Why:
|
||
# vLLM `Remove Sampler from Model Code` so vllm-ascend needs a patch to run in v0.8.4.
|
||
# How：
|
||
# Use vLLM 0.8.4 method tp patch it.
|
||
# Related PR (if no, explain why): 1. refused by vllm. 2. vllm doesn't support 3. prepare to submit....
|
||
# - https://github.com/vllm-project/vllm/pull/17084
|
||
# - https://github.com/vllm-project/vllm-ascend/pull/636
|
||
# Future Plan:
|
||
# Follow v0.8.4 version strategy.
|
||
#
|
||
# ** File: worker/patch_common/patch_metrics.py **
|
||
# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
|
||
# 1. `vllm.spec_decode.metrics.AsyncMetricsCollector.maybe_collect_rejsample_metrics`
|
||
# Why:
|
||
# There are cuda hard code (current_platform.is_cuda_alike()) in
|
||
# `AsyncMetricsCollector.maybe_collect_rejsample_metrics`
|
||
# How：
|
||
# Change to use `current_platform.Event` to determine whether to return None
|
||
# Related PR (if no, explain why): 1. refused by vllm. 2. vllm doesn't support 3. prepare to submit....
|
||
# https://github.com/vllm-project/vllm/pull/14411
|
||
# Future Plan:
|
||
# Revert it when the related pr is merged in vllm.
|
||
#
|
||
# ** File: worker/patch_common/patch_minicpm.py **
|
||
# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
|
||
# 1. `vllm.model_executor.models.minicpm.MiniCPMAttention.forward`
|
||
# Why:
|
||
# The forward func of MiniCPMAttention in vllm do a datatype convert
|
||
# (original datatype --> float32) to ensure the precision on cuda.
|
||
# However float32 is not supported in cann rope op, thus we keep this patch
|
||
# How：
|
||
# Removed the dtype convert operations in forward
|
||
# Related PR (if no, explain why): 1. refused by vllm. 2. vllm doesn't support 3. prepare to submit....
|
||
# NO, only for npu due to rope op.
|
||
# Future Plan:
|
||
# Keep this patch in vllm-ascend.
|
||
#
|
||
# ** File: worker/patch_common/patch_multi_step_worker.py **
|
||
# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
|
||
# 1. `vllm.spec_decode.multi_step_worker.MultiStepWorker.sampler_output`
|
||
# Why:
|
||
# There are cuda hard code (current_platform.is_cuda_alike()) in
|
||
# `MultiStepWorker.sampler_output`, and we need to use the patched `TP1DraftModelRunner` in it.
|
||
# How：
|
||
# Make speculative decoding extensible to different backends.
|
||
# - support attention metadata register to the set supported spec decode
|
||
# - offer a api in platform to determine whether spec decode is supported,
|
||
# and deprecate is_cuda_alike in it.
|
||
# Related PR (if no, explain why): 1. refused by vllm. 2. vllm doesn't support 3. prepare to submit....
|
||
# - https://github.com/vllm-project/vllm/pull/15195
|
||
# - https://github.com/vllm-project/vllm-ascend/pull/395
|
||
# Future Plan:
|
||
# Revert it when the related pr is merged in vllm and vllm-ascend.
|
||
#
|
||
# 2. `vllm.spec_decode.multi_step_worker.MultiStepWorker.set_include_gpu_probs_tensor` and
|
||
# `vllm.spec_decode.multi_step_worker.MultiStepWorker.set_should_modify_greedy_probs_inplace`
|
||
# Why:
|
||
# vLLM `Remove Sampler from Model Code` so vllm-ascend needs adapt to this change.
|
||
# How：
|
||
# Use vLLM 0.8.4 method to patch it.
|
||
# Related PR (if no, explain why): 1. refused by vllm. 2. vllm doesn't support 3. prepare to submit....
|
||
# - https://github.com/vllm-project/vllm/pull/15195
|
||
# - https://github.com/vllm-project/vllm-ascend/pull/395
|
||
# Future Plan:
|
||
# Remove it when we identify the reasons clearly.
|
||
#
|
||
# ** File: worker/patch_common/patch_spec_decode_worker.py **
|
||
# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
|
||
# 1. `vllm.spec_decode.spec_decode_worker.SpecDecodeWorker.create_worker`
|
||
# Why:
|
||
# We need to use the patched `TP1DraftModelRunner` in `SpecDecodeWorker.create_worker`.
|
||
# The mainly reason to overwrite `TP1DraftModelRunner`is the hard code of
|
||
# `FlashAttentionMetadata`
|
||
# How：
|
||
# ditto
|
||
# Related PR (if no, explain why): 1. refused by vllm. 2. vllm doesn't support 3. prepare to submit....
|
||
# - https://github.com/vllm-project/vllm/pull/15195
|
||
# - https://github.com/vllm-project/vllm-ascend/pull/395
|
||
# Future Plan:
|
||
# Revert it when the related pr is merged in vllm and vllm-ascend.
|
||
#
|
||
# ** File: worker/patch_0_8_4/patch_tritonplaceholder.py **
|
||
# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
|
||
# 1. `triton` Module
|
||
# Why:
|
||
# Triton is not supported on npu currently, importing triton will break vllm-ascend
|
||
# How：
|
||
# ditto
|
||
# Related PR (if no, explain why): 1. refused by vllm. 2. vllm doesn't support 3. prepare to submit....
|
||
# TritonPlaceholder is only available in vllm>0.8.4
|
||
# Future Plan:
|
||
# Revert it when branch main doesn't maintain v0.8.4.
|