EngineX/xc-llm-ascend
3
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity

[Misc] Move lora patch file into lora module (#2797)

Browse Source 
Cleanup useless file in patch module. Update the lora support list is OK
in vLLM Ascend, no need to patch vLLM


- vLLM version: v0.10.1.1
- vLLM main:
f4962a6d55

Signed-off-by: wangxiyuan <wangxiyuan1007@gmail.com>
This commit is contained in:
wangxiyuan
2025-09-08 21:42:12 +08:00
committed by GitHub
parent 85d989a3b9
commit 7d6d9449a8
10 changed files with 64 additions and 72 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
vllm_ascend/patch/worker/patch_common/__init__.py
View File

@@ -17,7 +17,4 @@
import vllm_ascend.patch.worker.patch_common.patch_distributed # noqa
import vllm_ascend.patch.worker.patch_common.patch_logits # noqa
import vllm_ascend.patch.worker.patch_common.patch_lora # noqa
import vllm_ascend.patch.worker.patch_common.patch_lora_embedding # noqa
import vllm_ascend.patch.worker.patch_common.patch_lora_linear # noqa
import vllm_ascend.patch.worker.patch_common.patch_minicpm # noqa

15
vllm_ascend/patch/worker/patch_common/patch_lora.py
View File

@@ -1,15 +0,0 @@
import vllm
from vllm.lora.utils import _all_lora_classes
from vllm_ascend.patch.worker.patch_common.patch_lora_embedding import \
AscendVocabParallelEmbeddingWithLoRA
from vllm_ascend.patch.worker.patch_common.patch_lora_linear import (
AscendColumnParallelLinearWithLoRA,
AscendMergedColumnParallelLinearWithLoRA, AscendRowParallelLinearWithLoRA)
_all_lora_classes.add(AscendRowParallelLinearWithLoRA)
_all_lora_classes.add(AscendColumnParallelLinearWithLoRA)
_all_lora_classes.add(AscendMergedColumnParallelLinearWithLoRA)
_all_lora_classes.add(AscendVocabParallelEmbeddingWithLoRA)
vllm.lora.utils._all_lora_classes = _all_lora_classes

22
vllm_ascend/patch/worker/patch_common/patch_lora_embedding.py
View File

@@ -1,22 +0,0 @@
from typing import Optional
from torch import nn
from transformers import PretrainedConfig
from vllm.config import LoRAConfig
from vllm.lora.layers import VocabParallelEmbeddingWithLoRA
from vllm_ascend.ops.vocab_parallel_embedding import \
AscendVocabParallelEmbedding
class AscendVocabParallelEmbeddingWithLoRA(VocabParallelEmbeddingWithLoRA):
@classmethod
def can_replace_layer(
cls,
source_layer: nn.Module,
lora_config: LoRAConfig,
packed_modules_list: list,
model_config: Optional[PretrainedConfig],
) -> bool:
return type(source_layer) is AscendVocabParallelEmbedding

52
vllm_ascend/patch/worker/patch_common/patch_lora_linear.py
View File

@@ -1,52 +0,0 @@
from typing import Optional
from torch import nn
from transformers import PretrainedConfig
from vllm.config import LoRAConfig
from vllm.lora.layers import (ColumnParallelLinearWithLoRA,
MergedColumnParallelLinearWithLoRA,
RowParallelLinearWithLoRA)
from vllm_ascend.ops.linear import (AscendColumnParallelLinear,
AscendMergedColumnParallelLinear,
AscendRowParallelLinear)
class AscendRowParallelLinearWithLoRA(RowParallelLinearWithLoRA):
@classmethod
def can_replace_layer(
cls,
source_layer: nn.Module,
lora_config: LoRAConfig,
packed_modules_list: list,
model_config: Optional[PretrainedConfig],
) -> bool:
return type(source_layer) is AscendRowParallelLinear
class AscendColumnParallelLinearWithLoRA(ColumnParallelLinearWithLoRA):
@classmethod
def can_replace_layer(
cls,
source_layer: nn.Module,
lora_config: LoRAConfig,
packed_modules_list: list,
model_config: Optional[PretrainedConfig],
) -> bool:
return type(source_layer) is AscendColumnParallelLinear
class AscendMergedColumnParallelLinearWithLoRA(
MergedColumnParallelLinearWithLoRA):
@classmethod
def can_replace_layer(
cls,
source_layer: nn.Module,
lora_config: LoRAConfig,
packed_modules_list: list,
model_config: Optional[PretrainedConfig],
) -> bool:
return type(source_layer) is AscendMergedColumnParallelLinear