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

Drop 0.11.0 support (#4377)

Browse Source 
There is a lot hack code for v0.11.0, which makes the code hard to
upgrade to newer vLLM version. Since v0.11.0 will release soon. Let's
drop v0.11.0 support first. Then we'll upgrade to v0.11.2 soon.


- vLLM version: v0.11.0
- vLLM main:
2918c1b49c

Signed-off-by: wangxiyuan <wangxiyuan1007@gmail.com>
This commit is contained in:
wangxiyuan
2025-11-24 17:08:20 +08:00
committed by GitHub
parent 41ddb06554
commit a1f142b7ad
80 changed files with 467 additions and 1755 deletions
Download Patch File Download Diff File Expand all files Collapse all files

14
vllm_ascend/model_loader/netloader/netloader.py
View File

@@ -29,18 +29,12 @@ from vllm.model_executor.model_loader.base_loader import BaseModelLoader
from vllm.model_executor.model_loader.default_loader import DefaultModelLoader
from vllm.model_executor.model_loader.utils import (
initialize_model, process_weights_after_loading)
from vllm_ascend.utils import vllm_version_is
from vllm.utils.torch_utils import set_default_torch_dtype
from .interaction.elastic import ElasticServer
from .load import elastic_load
from .utils import find_free_port, is_valid_path_prefix
if vllm_version_is("0.11.0"):
from vllm.model_executor.model_loader.utils import set_default_torch_dtype
else:
from vllm.utils.torch_utils import set_default_torch_dtype
@register_model_loader("netloader")
class ModelNetLoaderElastic(BaseModelLoader):
@@ -207,10 +201,8 @@ class ModelNetLoaderElastic(BaseModelLoader):
if model is not None and (
(self.listen_port and self.listen_port in range(1024, 65535)) or
(self.listen_port is None)):
if vllm_version_is("0.11.0"):
from vllm.utils import get_ip
else:
from vllm.utils.network_utils import get_ip
from vllm.utils.network_utils import get_ip
driver_ip = get_ip()
if driver_ip == '0.0.0.0':