diff --git a/vllm-0.6.6.post1.tar.gz b/vllm-0.7.3.tar.gz
similarity index 38%
rename from vllm-0.6.6.post1.tar.gz
rename to vllm-0.7.3.tar.gz
index f0884126b4a05cb81c1825f94ee1b1ef5c6834a9..d58972a9f9feba5a553789bb581fc98945998cc5 100644
Binary files a/vllm-0.6.6.post1.tar.gz and b/vllm-0.7.3.tar.gz differ
diff --git a/vllm.spec b/vllm.spec
index cc5f9d67e618e67d32aae5998d94b5190f926481..df5f0c2d9ed4f41bdc5233ac3a70056d90d9fc33 100644
--- a/vllm.spec
+++ b/vllm.spec
@@ -1,9 +1,10 @@
+%global _pip_huawei_install pip install --trusted-host mirrors.huaweicloud.com -i https://mirrors.huaweicloud.com/repository/pypi/simple
 %global _name vllm
 %global _description A high-throughput and memory-efficient inference and serving engine for LLMs
 
 Name:       vllm
-Version:    0.6.6.post1
-Release:    2
+Version:    0.7.3
+Release:    1
 Summary:    Powerful engine for LLMs
 License:    (Apache-2.0 AND BSD-3-Clause) OR BSD-3-CLause
 URL:        https://github.com/vllm-project/vllm
@@ -16,8 +17,8 @@ BuildArch:  noarch
 
 %package -n python3-%{_name}
 Summary:    %{summary}
-Buildrequires:  cmake python3-pip python3-devel python3-setuptools python3-pytest
-Buildrequires:  python3-setuptools_scm python3-wheel python3-grpcio
+Buildrequires:  cmake python3-pip python3-devel python3-pytest
+Buildrequires:  python3-wheel python3-grpcio
 Buildrequires:  python3-pytorch
 %{?python_provide:%python_provide python3-%{_name}}
 
@@ -28,6 +29,8 @@ Buildrequires:  python3-pytorch
 
 %prep
 %autosetup -n %{name}-%{version} -N
+%{_pip_huawei_install} install setuptools==68.0.0
+%{_pip_huawei_install} install setuptools-scm==6.0.0
 
 %build
 export SETUPTOOLS_SCM_PRETEND_VERSION=%{version}
@@ -69,6 +72,9 @@ mv %{buildroot}/filelist.lst .
 %files -n python3-%{_name} -f filelist.lst
 
 %changelog
+* Tue Apr 8 2025 renwenjie<renwenjie5@huawei.com> - 0.7.3-1
+- Change the baseline version to 0.7.3
+
 * Thu Mar 27 2025 renwenjie<renwenjie5@huawei.com> - 0.6.6.post1-2
 - solve 'no module named vllm._version' problem and adapted for openeuler22.03