From ba3d8aae943271221ff7b1ae74551939fd935ee5 Mon Sep 17 00:00:00 2001 From: Mengqing Cao Date: Sun, 27 Apr 2025 11:27:24 +0800 Subject: [PATCH] [Model][MiniCPM] support MiniCPM (#645) ### What this PR does / why we need it? This pr support minicpm in branch main. see https://github.com/vllm-project/vllm-ascend/pull/164 ### How was this patch tested? test locally with minicpm --------- Signed-off-by: MengqingCao --- vllm_ascend/patch/__init__.py | 28 ++++++++ .../patch/worker/patch_0_8_4/__init__.py | 1 + .../patch_0_8_4/patch_tritonplaceholder.py | 68 +++++++++++++++++++ .../patch/worker/patch_common/__init__.py | 1 + .../worker/patch_common/patch_minicpm.py | 36 ++++++++++ 5 files changed, 134 insertions(+) create mode 100644 vllm_ascend/patch/worker/patch_0_8_4/patch_tritonplaceholder.py create mode 100644 vllm_ascend/patch/worker/patch_common/patch_minicpm.py diff --git a/vllm_ascend/patch/__init__.py b/vllm_ascend/patch/__init__.py index 93c0b0b..ef0813a 100644 --- a/vllm_ascend/patch/__init__.py +++ b/vllm_ascend/patch/__init__.py @@ -96,6 +96,8 @@ # 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 ** @@ -125,6 +127,20 @@ # 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` @@ -156,3 +172,15 @@ # - 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. diff --git a/vllm_ascend/patch/worker/patch_0_8_4/__init__.py b/vllm_ascend/patch/worker/patch_0_8_4/__init__.py index 2465b3f..c822006 100644 --- a/vllm_ascend/patch/worker/patch_0_8_4/__init__.py +++ b/vllm_ascend/patch/worker/patch_0_8_4/__init__.py @@ -16,3 +16,4 @@ # import vllm_ascend.patch.worker.patch_0_8_4.patch_metrics # noqa +import vllm_ascend.patch.worker.patch_0_8_4.patch_tritonplaceholder # noqa diff --git a/vllm_ascend/patch/worker/patch_0_8_4/patch_tritonplaceholder.py b/vllm_ascend/patch/worker/patch_0_8_4/patch_tritonplaceholder.py new file mode 100644 index 0000000..cd32cd1 --- /dev/null +++ b/vllm_ascend/patch/worker/patch_0_8_4/patch_tritonplaceholder.py @@ -0,0 +1,68 @@ +# +# 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/triton_utils/importing.py +# + +import sys +import types +from importlib.util import find_spec + +from vllm.logger import logger + +HAS_TRITON = ( + find_spec("triton") is not None + or find_spec("pytorch-triton-xpu") is not None # Not compatible +) + +if not HAS_TRITON: + logger.info("Triton not installed or not compatible; certain GPU-related" + " functions will not be available.") + + class TritonPlaceholder(types.ModuleType): + + def __init__(self): + super().__init__("triton") + self.jit = self._dummy_decorator("jit") + self.autotune = self._dummy_decorator("autotune") + self.heuristics = self._dummy_decorator("heuristics") + self.language = TritonLanguagePlaceholder() + logger.warning_once( + "Triton is not installed. Using dummy decorators. " + "Install it via `pip install triton` to enable kernel" + " compilation.") + + def _dummy_decorator(self, name): + + def decorator(func=None, **kwargs): + if func is None: + return lambda f: f + return func + + return decorator + + class TritonLanguagePlaceholder(types.ModuleType): + + def __init__(self): + super().__init__("triton.language") + self.constexpr = None + self.dtype = None + + sys.modules['triton'] = TritonPlaceholder() + sys.modules['triton.language'] = TritonLanguagePlaceholder() + +if 'triton' in sys.modules: + logger.info("Triton module has been replaced with a placeholder.") diff --git a/vllm_ascend/patch/worker/patch_common/__init__.py b/vllm_ascend/patch/worker/patch_common/__init__.py index 590074f..2ba9917 100644 --- a/vllm_ascend/patch/worker/patch_common/__init__.py +++ b/vllm_ascend/patch/worker/patch_common/__init__.py @@ -16,5 +16,6 @@ # import vllm_ascend.patch.worker.patch_common.patch_metrics # noqa +import vllm_ascend.patch.worker.patch_common.patch_minicpm # noqa import vllm_ascend.patch.worker.patch_common.patch_multi_step_worker # noqa import vllm_ascend.patch.worker.patch_common.patch_spec_decode_worker # noqa diff --git a/vllm_ascend/patch/worker/patch_common/patch_minicpm.py b/vllm_ascend/patch/worker/patch_common/patch_minicpm.py new file mode 100644 index 0000000..663a08a --- /dev/null +++ b/vllm_ascend/patch/worker/patch_common/patch_minicpm.py @@ -0,0 +1,36 @@ +# +# 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. +# + +import torch +from vllm.model_executor.models.minicpm import MiniCPMAttention + + +def forward( + self, + positions: torch.Tensor, + hidden_states: torch.Tensor, +) -> torch.Tensor: + qkv, _ = self.qkv_proj(hidden_states) + q, k, v = qkv.split([self.q_size, self.kv_size, self.kv_size], dim=-1) + q, k = self.rotary_emb(positions, q, k) + attn_output = self.attn(q, k, v) + output, _ = self.o_proj(attn_output) + return output + + +# The type conversion in the forward function is deleted to support the rope operator. +MiniCPMAttention.forward = forward