Use vllm-gaudi:1.22.0 image for AudioQnA and ChatQnA (#2266)

ZePan110 · web-flow · commit 5d8cffd40eb2 · 2025-09-16T14:39:01.000+08:00
Signed-off-by: ZePan110 &lt;ze.pan@intel.com&gt;
diff --git a/AudioQnA/docker_compose/intel/hpu/gaudi/compose.yaml b/AudioQnA/docker_compose/intel/hpu/gaudi/compose.yaml
@@ -35,7 +35,7 @@ services:
       - SYS_NICE
     restart: unless-stopped
   vllm-service:
-    image: ${REGISTRY:-opea}/vllm-gaudi:${TAG:-latest}
+    image: opea/vllm-gaudi:1.22.0
     container_name: vllm-gaudi-service
     ports:
       - ${LLM_SERVER_PORT:-3006}:80
diff --git a/AudioQnA/docker_image_build/build.yaml b/AudioQnA/docker_image_build/build.yaml
@@ -102,12 +102,6 @@ services:
       dockerfile: comps/third_parties/gpt-sovits/src/Dockerfile
     extends: audioqna
     image: ${REGISTRY:-opea}/gpt-sovits:${TAG:-latest}
-  vllm-gaudi:
-    build:
-      context: vllm-fork
-      dockerfile: Dockerfile.hpu
-    extends: audioqna
-    image: ${REGISTRY:-opea}/vllm-gaudi:${TAG:-latest}
   vllm-rocm:
     build:
       context: GenAIComps
diff --git a/AudioQnA/tests/test_compose_multilang_on_xeon.sh b/AudioQnA/tests/test_compose_multilang_on_xeon.sh
@@ -35,6 +35,7 @@ function build_docker_images() {
 function start_services() {
     cd $WORKPATH/docker_compose/intel/cpu/xeon/
     export host_ip=${ip_address}
+    export no_proxy="localhost,127.0.0.1,$ip_address"
     source set_env.sh
     # sed -i "s/backend_address/$ip_address/g" $WORKPATH/ui/svelte/.env
 
diff --git a/AudioQnA/tests/test_compose_on_epyc.sh b/AudioQnA/tests/test_compose_on_epyc.sh
@@ -37,6 +37,7 @@ function build_docker_images() {
 function start_services() {
 	cd $WORKPATH/docker_compose/amd/cpu/epyc/
 	export host_ip=${ip_address}
+	export no_proxy="localhost,127.0.0.1,$ip_address"
 	source set_env.sh
 	# sed -i "s/backend_address/$ip_address/g" $WORKPATH/ui/svelte/.env
 
diff --git a/AudioQnA/tests/test_compose_on_gaudi.sh b/AudioQnA/tests/test_compose_on_gaudi.sh
@@ -25,14 +25,8 @@ function build_docker_images() {
     docker build --no-cache -t ${REGISTRY}/comps-base:${TAG} --build-arg https_proxy=$https_proxy --build-arg http_proxy=$http_proxy -f Dockerfile .
     popd && sleep 1s
 
-    git clone https://github.com/HabanaAI/vllm-fork.git
-    cd vllm-fork/
-    VLLM_FORK_VER=v0.6.6.post1+Gaudi-1.20.0
-    echo "Check out vLLM tag ${VLLM_FORK_VER}"
-    git checkout ${VLLM_FORK_VER} &> /dev/null && cd ../
-
     echo "Build all the images with --no-cache, check docker_image_build.log for details..."
-    service_list="audioqna audioqna-ui whisper-gaudi speecht5-gaudi vllm-gaudi"
+    service_list="audioqna audioqna-ui whisper-gaudi speecht5-gaudi"
     docker compose -f build.yaml build ${service_list} --no-cache > ${LOG_PATH}/docker_image_build.log
 
     docker images && sleep 1s
@@ -41,6 +35,7 @@ function build_docker_images() {
 function start_services() {
     cd $WORKPATH/docker_compose/intel/hpu/gaudi
     export host_ip=${ip_address}
+    export no_proxy="localhost,127.0.0.1,$ip_address"
     source set_env.sh
     # sed -i "s/backend_address/$ip_address/g" $WORKPATH/ui/svelte/.env
 
diff --git a/AudioQnA/tests/test_compose_on_rocm.sh b/AudioQnA/tests/test_compose_on_rocm.sh
@@ -35,6 +35,7 @@ function build_docker_images() {
 
 function start_services() {
     cd $WORKPATH/docker_compose/amd/gpu/rocm/
+    export no_proxy="localhost,127.0.0.1,$ip_address"
     source set_env.sh
     # Start Docker Containers
     docker compose up -d > ${LOG_PATH}/start_services_with_compose.log
diff --git a/AudioQnA/tests/test_compose_on_xeon.sh b/AudioQnA/tests/test_compose_on_xeon.sh
@@ -35,6 +35,7 @@ function build_docker_images() {
 function start_services() {
     cd $WORKPATH/docker_compose/intel/cpu/xeon/
     export host_ip=${ip_address}
+    export no_proxy="localhost,127.0.0.1,$ip_address"
     source set_env.sh
     # sed -i "s/backend_address/$ip_address/g" $WORKPATH/ui/svelte/.env
 
diff --git a/AudioQnA/tests/test_compose_openeuler_on_xeon.sh b/AudioQnA/tests/test_compose_openeuler_on_xeon.sh
@@ -35,6 +35,7 @@ function build_docker_images() {
 function start_services() {
     cd $WORKPATH/docker_compose/intel/cpu/xeon/
     export host_ip=${ip_address}
+    export no_proxy="localhost,127.0.0.1,$ip_address"
     source set_env.sh
     # sed -i "s/backend_address/$ip_address/g" $WORKPATH/ui/svelte/.env
 
diff --git a/AudioQnA/tests/test_compose_tgi_on_epyc.sh b/AudioQnA/tests/test_compose_tgi_on_epyc.sh
@@ -38,6 +38,7 @@ function build_docker_images() {
 function start_services() {
     echo $WORKPATH
     cd $WORKPATH/docker_compose/amd/cpu/epyc/
+    export no_proxy="localhost,127.0.0.1,$ip_address"
     source set_env.sh
     # Start Docker Containers
     docker compose -f compose_tgi.yaml up -d > ${LOG_PATH}/start_services_with_compose.log
diff --git a/AudioQnA/tests/test_compose_tgi_on_gaudi.sh b/AudioQnA/tests/test_compose_tgi_on_gaudi.sh
@@ -35,6 +35,7 @@ function build_docker_images() {
 function start_services() {
     cd $WORKPATH/docker_compose/intel/hpu/gaudi
     export host_ip=${ip_address}
+    export no_proxy="localhost,127.0.0.1,$ip_address"
     source set_env.sh
     # Start Docker Containers
     docker compose -f compose_tgi.yaml up -d > ${LOG_PATH}/start_services_with_compose.log
diff --git a/AudioQnA/tests/test_compose_tgi_on_xeon.sh b/AudioQnA/tests/test_compose_tgi_on_xeon.sh
@@ -35,6 +35,7 @@ function build_docker_images() {
 function start_services() {
     cd $WORKPATH/docker_compose/intel/cpu/xeon/
     export host_ip=${ip_address}
+    export no_proxy="localhost,127.0.0.1,$ip_address"
     source set_env.sh
     # Start Docker Containers
     docker compose -f compose_tgi.yaml up -d > ${LOG_PATH}/start_services_with_compose.log
diff --git a/AudioQnA/tests/test_compose_vllm_on_rocm.sh b/AudioQnA/tests/test_compose_vllm_on_rocm.sh
@@ -33,6 +33,7 @@ function build_docker_images() {
 
 function start_services() {
     cd $WORKPATH/docker_compose/amd/gpu/rocm/
+    export no_proxy="localhost,127.0.0.1,$ip_address"
     source set_env_vllm.sh
     sed -i "s/backend_address/$ip_address/g" $WORKPATH/ui/svelte/.env
 
diff --git a/ChatQnA/docker_compose/intel/hpu/gaudi/compose.yaml b/ChatQnA/docker_compose/intel/hpu/gaudi/compose.yaml
@@ -92,7 +92,7 @@ services:
       MAX_WARMUP_SEQUENCE_LENGTH: 512
     command: --model-id ${RERANK_MODEL_ID} --auto-truncate
   vllm-service:
-    image: ${REGISTRY:-opea}/vllm-gaudi:${TAG:-latest}
+    image: ${REGISTRY:-opea}/vllm-gaudi:1.22.0
     container_name: vllm-gaudi-server
     ports:
       - "8007:80"
diff --git a/ChatQnA/docker_compose/intel/hpu/gaudi/compose_faqgen.yaml b/ChatQnA/docker_compose/intel/hpu/gaudi/compose_faqgen.yaml
@@ -85,7 +85,7 @@ services:
       MAX_WARMUP_SEQUENCE_LENGTH: 512
     command: --model-id ${RERANK_MODEL_ID} --auto-truncate
   vllm-service:
-    image: ${REGISTRY:-opea}/vllm-gaudi:${TAG:-latest}
+    image: opea/vllm-gaudi:1.22.0
     container_name: vllm-gaudi-server
     ports:
       - ${LLM_ENDPOINT_PORT:-8007}:80
diff --git a/ChatQnA/docker_compose/intel/hpu/gaudi/compose_guardrails.yaml b/ChatQnA/docker_compose/intel/hpu/gaudi/compose_guardrails.yaml
@@ -32,7 +32,7 @@ services:
       retries: 50
     restart: unless-stopped
   vllm-guardrails-service:
-    image: ${REGISTRY:-opea}/vllm-gaudi:${TAG:-latest}
+    image: opea/vllm-gaudi:1.22.0
     container_name: vllm-guardrails-server
     ports:
       - "8088:80"
@@ -130,7 +130,7 @@ services:
       MAX_WARMUP_SEQUENCE_LENGTH: 512
     command: --model-id ${RERANK_MODEL_ID} --auto-truncate
   vllm-service:
-    image: ${REGISTRY:-opea}/vllm-gaudi:${TAG:-latest}
+    image: ${REGISTRY:-opea}/vllm-gaudi:1.22.0
     container_name: vllm-gaudi-server
     ports:
       - "8008:80"
diff --git a/ChatQnA/docker_compose/intel/hpu/gaudi/compose_without_rerank.yaml b/ChatQnA/docker_compose/intel/hpu/gaudi/compose_without_rerank.yaml
@@ -65,7 +65,7 @@ services:
       RETRIEVER_COMPONENT_NAME: "OPEA_RETRIEVER_REDIS"
     restart: unless-stopped
   vllm-service:
-    image: ${REGISTRY:-opea}/vllm-gaudi:${TAG:-latest}
+    image: opea/vllm-gaudi:1.22.0
     container_name: vllm-gaudi-server
     ports:
       - "8007:80"
diff --git a/ChatQnA/docker_image_build/build.yaml b/ChatQnA/docker_image_build/build.yaml
@@ -114,12 +114,6 @@ services:
       context: GenAIComps
       dockerfile: comps/third_parties/vllm/src/Dockerfile.amd_gpu
     image: ${REGISTRY:-opea}/vllm-rocm:${TAG:-latest}
-  vllm-gaudi:
-    build:
-      context: vllm-fork
-      dockerfile: Dockerfile.hpu
-    extends: chatqna
-    image: ${REGISTRY:-opea}/vllm-gaudi:${TAG:-latest}
   nginx:
     build:
       context: GenAIComps
diff --git a/ChatQnA/tests/test_compose_faqgen_on_gaudi.sh b/ChatQnA/tests/test_compose_faqgen_on_gaudi.sh
@@ -23,19 +23,17 @@ function build_docker_images() {
     echo "GenAIComps test commit is $(git rev-parse HEAD)"
     docker build --no-cache -t ${REGISTRY}/comps-base:${TAG} --build-arg https_proxy=$https_proxy --build-arg http_proxy=$http_proxy -f Dockerfile .
     popd && sleep 1s
-    git clone https://github.com/HabanaAI/vllm-fork.git && cd vllm-fork
-    VLLM_FORK_VER=v0.6.6.post1+Gaudi-1.20.0
-    git checkout ${VLLM_FORK_VER} &> /dev/null && cd ../
 
     echo "Build all the images with --no-cache, check docker_image_build.log for details..."
-    service_list="chatqna chatqna-ui dataprep retriever llm-faqgen vllm-gaudi nginx"
+    service_list="chatqna chatqna-ui dataprep retriever llm-faqgen nginx"
     docker compose -f build.yaml build ${service_list} --no-cache > ${LOG_PATH}/docker_image_build.log
 
     docker images && sleep 1s
 }
 
 function start_services() {
     cd $WORKPATH/docker_compose/intel/hpu/gaudi
+    export no_proxy="localhost,127.0.0.1,$ip_address"
     source set_env_faqgen.sh
 
     # Start Docker Containers
diff --git a/ChatQnA/tests/test_compose_faqgen_tgi_on_gaudi.sh b/ChatQnA/tests/test_compose_faqgen_tgi_on_gaudi.sh
@@ -34,6 +34,7 @@ function build_docker_images() {
 function start_services() {
     cd $WORKPATH/docker_compose/intel/hpu/gaudi
     export FAQGen_COMPONENT_NAME="OpeaFaqGenTgi"
+    export no_proxy="localhost,127.0.0.1,$ip_address"
     source set_env_faqgen.sh
 
     # Start Docker Containers
diff --git a/ChatQnA/tests/test_compose_guardrails_on_gaudi.sh b/ChatQnA/tests/test_compose_guardrails_on_gaudi.sh
@@ -23,12 +23,9 @@ function build_docker_images() {
     echo "GenAIComps test commit is $(git rev-parse HEAD)"
     docker build --no-cache -t ${REGISTRY}/comps-base:${TAG} --build-arg https_proxy=$https_proxy --build-arg http_proxy=$http_proxy -f Dockerfile .
     popd && sleep 1s
-    git clone https://github.com/HabanaAI/vllm-fork.git && cd vllm-fork
-    VLLM_FORK_VER=v0.6.6.post1+Gaudi-1.20.0
-    git checkout ${VLLM_FORK_VER} &> /dev/null && cd ../
 
     echo "Build all the images with --no-cache, check docker_image_build.log for details..."
-    service_list="chatqna chatqna-ui dataprep retriever vllm-gaudi guardrails nginx"
+    service_list="chatqna chatqna-ui dataprep retriever guardrails nginx"
     docker compose -f build.yaml build ${service_list} --no-cache > ${LOG_PATH}/docker_image_build.log
 
     docker images && sleep 1s
diff --git a/ChatQnA/tests/test_compose_mariadb_on_xeon.sh b/ChatQnA/tests/test_compose_mariadb_on_xeon.sh
@@ -34,6 +34,7 @@ function build_docker_images() {
 function start_services() {
     cd $WORKPATH/docker_compose/intel/cpu/xeon
     export MARIADB_PASSWORD="test"
+    export no_proxy="localhost,127.0.0.1,$ip_address"
     source set_env_mariadb.sh
 
     # Start Docker Containers
diff --git a/ChatQnA/tests/test_compose_milvus_on_xeon.sh b/ChatQnA/tests/test_compose_milvus_on_xeon.sh
@@ -34,6 +34,7 @@ function build_docker_images() {
 function start_services() {
     cd $WORKPATH/docker_compose/intel/cpu/xeon/
     export LOGFLAG=true
+    export no_proxy="localhost,127.0.0.1,$ip_address"
     source set_env.sh
 
     # Start Docker Containers
diff --git a/ChatQnA/tests/test_compose_on_gaudi.sh b/ChatQnA/tests/test_compose_on_gaudi.sh
@@ -23,12 +23,9 @@ function build_docker_images() {
     echo "GenAIComps test commit is $(git rev-parse HEAD)"
     docker build --no-cache -t ${REGISTRY}/comps-base:${TAG} --build-arg https_proxy=$https_proxy --build-arg http_proxy=$http_proxy -f Dockerfile .
     popd && sleep 1s
-    git clone https://github.com/HabanaAI/vllm-fork.git && cd vllm-fork
-    VLLM_FORK_VER=v0.6.6.post1+Gaudi-1.20.0
-    git checkout ${VLLM_FORK_VER} &> /dev/null && cd ../
 
     echo "Build all the images with --no-cache, check docker_image_build.log for details..."
-    service_list="chatqna chatqna-ui dataprep retriever vllm-gaudi nginx"
+    service_list="chatqna chatqna-ui dataprep retriever nginx"
     docker compose -f build.yaml build ${service_list} --no-cache > ${LOG_PATH}/docker_image_build.log
 
     docker images && sleep 1s
diff --git a/ChatQnA/tests/test_compose_on_xeon.sh b/ChatQnA/tests/test_compose_on_xeon.sh
@@ -33,7 +33,7 @@ function build_docker_images() {
 
 function start_services() {
     cd $WORKPATH/docker_compose/intel/cpu/xeon
-
+    export no_proxy="localhost,127.0.0.1,$ip_address"
     source set_env.sh
 
     # Start Docker Containers
diff --git a/ChatQnA/tests/test_compose_openeuler_on_xeon.sh b/ChatQnA/tests/test_compose_openeuler_on_xeon.sh
@@ -33,7 +33,7 @@ function build_docker_images() {
 
 function start_services() {
     cd $WORKPATH/docker_compose/intel/cpu/xeon
-
+    export no_proxy="localhost,127.0.0.1,$ip_address"
     source set_env.sh
 
     # Start Docker Containers
diff --git a/ChatQnA/tests/test_compose_qdrant_on_xeon.sh b/ChatQnA/tests/test_compose_qdrant_on_xeon.sh
@@ -33,7 +33,7 @@ function build_docker_images() {
 
 function start_services() {
     cd $WORKPATH/docker_compose/intel/cpu/xeon
-
+    export no_proxy="localhost,127.0.0.1,$ip_address"
     export INDEX_NAME="rag-qdrant"
     source set_env.sh
 
diff --git a/ChatQnA/tests/test_compose_tgi_on_gaudi.sh b/ChatQnA/tests/test_compose_tgi_on_gaudi.sh
@@ -35,6 +35,7 @@ function start_services() {
     export NON_INTERACTIVE=true
     export host_ip=${ip_address}
     export telemetry=yes
+    export no_proxy="localhost,127.0.0.1,$ip_address"
     source set_env.sh
 
     # Start Docker Containers
diff --git a/ChatQnA/tests/test_compose_tgi_on_xeon.sh b/ChatQnA/tests/test_compose_tgi_on_xeon.sh
@@ -32,7 +32,7 @@ function build_docker_images() {
 
 function start_services() {
     cd $WORKPATH/docker_compose/intel/cpu/xeon
-
+    export no_proxy="localhost,127.0.0.1,$ip_address"
     source set_env.sh
 
     # Start Docker Containers
diff --git a/ChatQnA/tests/test_compose_without_rerank_on_gaudi.sh b/ChatQnA/tests/test_compose_without_rerank_on_gaudi.sh
@@ -23,12 +23,9 @@ function build_docker_images() {
     echo "GenAIComps test commit is $(git rev-parse HEAD)"
     docker build --no-cache -t ${REGISTRY}/comps-base:${TAG} --build-arg https_proxy=$https_proxy --build-arg http_proxy=$http_proxy -f Dockerfile .
     popd && sleep 1s
-    git clone https://github.com/HabanaAI/vllm-fork.git && cd vllm-fork
-    VLLM_FORK_VER=v0.6.6.post1+Gaudi-1.20.0
-    git checkout ${VLLM_FORK_VER} &> /dev/null && cd ../
 
     echo "Build all the images with --no-cache, check docker_image_build.log for details..."
-    service_list="chatqna chatqna-ui dataprep retriever vllm-gaudi nginx"
+    service_list="chatqna chatqna-ui dataprep retriever nginx"
     docker compose -f build.yaml build ${service_list} --no-cache > ${LOG_PATH}/docker_image_build.log
 
     docker images && sleep 1s
diff --git a/ChatQnA/tests/test_compose_without_rerank_on_xeon.sh b/ChatQnA/tests/test_compose_without_rerank_on_xeon.sh
@@ -33,7 +33,7 @@ function build_docker_images() {
 
 function start_services() {
     cd $WORKPATH/docker_compose/intel/cpu/xeon
-
+    export no_proxy="localhost,127.0.0.1,$ip_address"
     source set_env.sh
 
     # Start Docker Containers
