2026-01-17 11:49:18 +08:00
|
|
|
#
|
|
|
|
|
# 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 torch_npu
|
|
|
|
|
|
2026-02-02 16:12:04 +08:00
|
|
|
from vllm_ascend.ops.mm_encoder_attention import AscendMMEncoderAttention
|
2026-01-17 11:49:18 +08:00
|
|
|
|
|
|
|
|
|
2026-02-02 16:12:04 +08:00
|
|
|
class AscendMMEncoderAttention310(AscendMMEncoderAttention):
|
2026-01-17 11:49:18 +08:00
|
|
|
def __init__(self, *args, **kwargs):
|
|
|
|
|
super().__init__(*args, **kwargs)
|
|
|
|
|
|
|
|
|
|
def forward_oot(
|
|
|
|
|
self,
|
|
|
|
|
query: torch.Tensor,
|
|
|
|
|
key: torch.Tensor,
|
|
|
|
|
value: torch.Tensor,
|
|
|
|
|
cu_seqlens: torch.Tensor | None = None,
|
|
|
|
|
max_seqlen: int | None = None,
|
|
|
|
|
**kwargs,
|
|
|
|
|
):
|
|
|
|
|
bsz, q_len = query.size()[:2]
|
|
|
|
|
kv_len = key.size(1)
|
2026-02-07 09:25:17 +08:00
|
|
|
query = query.view(bsz * q_len, self.num_heads, self.head_size)
|
|
|
|
|
key = key.view(bsz * kv_len, self.num_kv_heads, self.head_size)
|
|
|
|
|
value = value.view(bsz * kv_len, self.num_kv_heads, self.head_size)
|
2026-01-17 11:49:18 +08:00
|
|
|
|
|
|
|
|
if cu_seqlens is None:
|
2026-02-07 09:25:17 +08:00
|
|
|
seq_len = torch.tensor([q_len] * bsz, device="cpu", dtype=torch.int32)
|
|
|
|
|
else:
|
|
|
|
|
seq_len = torch.diff(cu_seqlens.to("cpu", dtype=torch.int32))
|
2026-01-24 20:34:29 +08:00
|
|
|
|
2026-02-07 09:25:17 +08:00
|
|
|
output = torch.empty_like(query)
|
2026-01-24 20:34:29 +08:00
|
|
|
torch_npu._npu_flash_attention_unpad(
|
2026-02-07 09:25:17 +08:00
|
|
|
query=query,
|
|
|
|
|
key=key,
|
|
|
|
|
value=value,
|
2026-01-24 20:34:29 +08:00
|
|
|
seq_len=seq_len,
|
|
|
|
|
scale_value=self.head_size**-0.5,
|
|
|
|
|
num_heads=self.num_heads,
|
|
|
|
|
num_kv_heads=self.num_kv_heads,
|
2026-02-07 09:25:17 +08:00
|
|
|
out=output,
|
2026-01-24 20:34:29 +08:00
|
|
|
)
|
|
|
|
|
|
2026-02-07 09:25:17 +08:00
|
|
|
output = output.view(bsz, -1, self.num_heads, self.head_size)
|
|
|
|
|
return output
|