94dd832815
### What this PR does / why we need it? 1. Move additional functionalities from fused_moe.py to common_fused_moe.py and remove fused_moe.py 2. Remove unnecessary custom classes from qwen3_moe.py, and it will be completely removed after we release vllm-ascend v0.11.0 ### Does this PR introduce _any_ user-facing change? No ### How was this patch tested? Qwen3-30B-A3B/Qwen3-30B-A3B-W8A8/DeepSeek-V3-W4A8-Pruing/deepseek-mtp/pangu-pro-moe-pruing: 1. Enable/Disable EP 3. Aclgraph & eager 4. SP - vLLM version: v0.11.0 --------- Signed-off-by: Pr0Wh1teGivee <calvin_zhu0210@outlook.com> Co-authored-by: weijinqian0 <12153182+weijinqian0@users.noreply.github.com>
58 lines
2.1 KiB
Python
58 lines
2.1 KiB
Python
#
|
|
# 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 torch
|
|
|
|
import vllm_ascend.ops.common_fused_moe # noqa
|
|
import vllm_ascend.ops.layernorm # noqa
|
|
import vllm_ascend.ops.register_custom_ops # noqa
|
|
import vllm_ascend.ops.vocab_parallel_embedding # noqa
|
|
from vllm_ascend.ops.activation import AscendQuickGELU, AscendSiluAndMul
|
|
from vllm_ascend.ops.rotary_embedding import (
|
|
AscendDeepseekScalingRotaryEmbedding, AscendRotaryEmbedding)
|
|
|
|
|
|
class dummyFusionOp:
|
|
default = None
|
|
|
|
def __init__(self, name=""):
|
|
self.name = name
|
|
|
|
|
|
def register_dummy_fusion_op() -> None:
|
|
torch.ops._C_ascend.rms_norm = dummyFusionOp(name="rms_norm")
|
|
torch.ops._C_ascend.fused_add_rms_norm = dummyFusionOp(
|
|
name="fused_add_rms_norm")
|
|
torch.ops._C_ascend.static_scaled_fp8_quant = dummyFusionOp(
|
|
name="static_scaled_fp8_quant")
|
|
torch.ops._C_ascend.dynamic_scaled_fp8_quant = dummyFusionOp(
|
|
name="dynamic_scaled_fp8_quant")
|
|
torch.ops._C_ascend.dynamic_per_token_scaled_fp8_quant = dummyFusionOp(
|
|
name="dynamic_per_token_scaled_fp8_quant")
|
|
torch.ops._C_ascend.rms_norm_static_fp8_quant = dummyFusionOp(
|
|
name="rms_norm_static_fp8_quant")
|
|
torch.ops._C_ascend.fused_add_rms_norm_static_fp8_quant = dummyFusionOp(
|
|
name="fused_add_rms_norm_static_fp8_quant")
|
|
torch.ops._C_ascend.rms_norm_dynamic_per_token_quant = dummyFusionOp(
|
|
name="rms_norm_dynamic_per_token_quant")
|
|
|
|
|
|
__all__ = [
|
|
"AscendQuickGELU", "AscendSiluAndMul", "AscendRotaryEmbedding",
|
|
"AscendDeepseekScalingRotaryEmbedding"
|
|
]
|