From 71282c7ee9ebaf27ae2a4af0ed3cf27d91aa57af Mon Sep 17 00:00:00 2001 From: Steve Zhang Date: Fri, 20 Sep 2024 09:19:46 +0800 Subject: [PATCH] 'ghcr.io/huggingface/text-generation-inference:sha-e4201f4-intel-cpu' is intel cpu optimized tgi image, we need to use this one for all xeon platform. (#444) Signed-off-by: zhlsunshine (cherry picked from commit c84ac4c74c9d5f1474a4966d0107830db2963d1d) --- helm-charts/common/tgi/values.yaml | 3 ++- microservices-connector/config/manifests/tgi.yaml | 2 +- microservices-connector/config/samples/ChatQnA/use_cases.md | 2 +- 3 files changed, 4 insertions(+), 3 deletions(-) diff --git a/helm-charts/common/tgi/values.yaml b/helm-charts/common/tgi/values.yaml index 805df10b..97ef2e59 100644 --- a/helm-charts/common/tgi/values.yaml +++ b/helm-charts/common/tgi/values.yaml @@ -26,7 +26,8 @@ image: repository: ghcr.io/huggingface/text-generation-inference pullPolicy: IfNotPresent # Overrides the image tag whose default is the chart appVersion. - tag: "2.2.0" + # `sha-e4201f4-intel-cpu` is the image tag for intel cpu optimized tgi image + tag: "sha-e4201f4-intel-cpu" # empty for CPU accelDevice: "" diff --git a/microservices-connector/config/manifests/tgi.yaml b/microservices-connector/config/manifests/tgi.yaml index aa1f4cec..cece9855 100644 --- a/microservices-connector/config/manifests/tgi.yaml +++ b/microservices-connector/config/manifests/tgi.yaml @@ -87,7 +87,7 @@ spec: optional: true securityContext: {} - image: "ghcr.io/huggingface/text-generation-inference:2.2.0" + image: "ghcr.io/huggingface/text-generation-inference:sha-e4201f4-intel-cpu" imagePullPolicy: IfNotPresent volumeMounts: - mountPath: /data diff --git a/microservices-connector/config/samples/ChatQnA/use_cases.md b/microservices-connector/config/samples/ChatQnA/use_cases.md index c8acc2c9..e18ae2f7 100644 --- a/microservices-connector/config/samples/ChatQnA/use_cases.md +++ b/microservices-connector/config/samples/ChatQnA/use_cases.md @@ -19,7 +19,7 @@ The ChatQnA uses the below prebuilt images if you choose a Xeon deployment - dataprep-redis: opea/dataprep-redis:latest - tei_xeon_service: ghcr.io/huggingface/text-embeddings-inference:cpu-1.5 - tei_embedding_service: ghcr.io/huggingface/text-embeddings-inference:cpu-1.5 -- tgi-service: ghcr.io/huggingface/text-generation-inference:2.2.0 +- tgi-service: ghcr.io/huggingface/text-generation-inference:sha-e4201f4-intel-cpu - redis-vector-db: redis/redis-stack:7.2.0-v9 Should you desire to use the Gaudi accelerator, two alternate images are used for the embedding and llm services.