1a1f9a6d89
### What this PR does / why we need it? This PR ports all the deepseek graph mode code and mtp code from v0.7.3 to the main branch --------- Signed-off-by: SidaoY <1024863041@qq.com> Signed-off-by: linfeng-yuan <1102311262@qq.com> Signed-off-by: Yizhou Liu <liuyizhou5@h-partners.com> Signed-off-by: mengwei805 <mengwei25@huawei.com> Signed-off-by: libaokui <libaokui@huawei.com> Signed-off-by: q00832892 <qiaoyang19@huawei.com> Signed-off-by: ganyi <pleaplusone.gy@gmail.com> Co-authored-by: SidaoY <1024863041@qq.com> Co-authored-by: linfeng-yuan <1102311262@qq.com> Co-authored-by: Yizhou Liu <liuyizhou5@h-partners.com> Co-authored-by: mengwei805 <mengwei25@huawei.com> Co-authored-by: libaokui <libaokui@huawei.com>
68 lines
2.7 KiB
Python
68 lines
2.7 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.
|
|
#
|
|
|
|
from typing import Tuple
|
|
|
|
import torch
|
|
from vllm.distributed import tensor_model_parallel_all_reduce
|
|
from vllm.model_executor.layers.vocab_parallel_embedding import \
|
|
VocabParallelEmbedding
|
|
|
|
|
|
def get_masked_input_and_mask(
|
|
input_: torch.Tensor, org_vocab_start_index: int,
|
|
org_vocab_end_index: int, num_org_vocab_padding: int,
|
|
added_vocab_start_index: int,
|
|
added_vocab_end_index: int) -> Tuple[torch.Tensor, torch.Tensor]:
|
|
# torch.compile will fuse all of the pointwise ops below
|
|
# into a single kernel, making it very fast
|
|
org_vocab_mask = (input_ >= org_vocab_start_index) & (input_ <
|
|
org_vocab_end_index)
|
|
added_vocab_mask = (input_ >= added_vocab_start_index) & (
|
|
input_ < added_vocab_end_index)
|
|
added_offset = added_vocab_start_index - (
|
|
org_vocab_end_index - org_vocab_start_index) - num_org_vocab_padding
|
|
valid_offset = (org_vocab_start_index *
|
|
org_vocab_mask) + (added_offset * added_vocab_mask)
|
|
vocab_mask = org_vocab_mask | added_vocab_mask
|
|
input_ = vocab_mask * (input_ - valid_offset)
|
|
return input_, ~vocab_mask
|
|
|
|
|
|
def vocab_parallel_embedding_forward(self, input_):
|
|
if self.tp_size > 1:
|
|
# Build the mask.
|
|
masked_input, input_mask = get_masked_input_and_mask(
|
|
input_, self.shard_indices.org_vocab_start_index,
|
|
self.shard_indices.org_vocab_end_index,
|
|
self.shard_indices.num_org_vocab_padding,
|
|
self.shard_indices.added_vocab_start_index,
|
|
self.shard_indices.added_vocab_end_index)
|
|
else:
|
|
masked_input = input_
|
|
# Get the embeddings.
|
|
output_parallel = self.quant_method.embedding(self, masked_input.long())
|
|
# Mask the output embedding.
|
|
if self.tp_size > 1:
|
|
output_parallel.masked_fill_(input_mask.unsqueeze(-1), 0)
|
|
# Reduce across all the model parallel GPUs.
|
|
output = tensor_model_parallel_all_reduce(output_parallel)
|
|
return output
|
|
|
|
|
|
VocabParallelEmbedding.forward = vocab_parallel_embedding_forward
|