helm: upgrade tgi chart

- upgrade tgi version to 2.1.0 - upgrade tgi-gaudi version to 2.0.1 - change default model to Intel/neural-chat-7b-v3-3 Signed-off-by: Lianhao Lu <[email protected]>
opea-project · Jul 18, 2024 · c3a1c1a · c3a1c1a
1 parent fbdb1da
commit c3a1c1a
Show file tree

Hide file tree

Showing 5 changed files with 15 additions and 17 deletions.
diff --git a/helm-charts/common/tgi/Chart.yaml b/helm-charts/common/tgi/Chart.yaml
@@ -7,4 +7,4 @@ description: The Helm chart for HuggingFace Text Generation Inference Server
 type: application
 version: 0.8.0
 # The HF TGI version
-appVersion: "1.4"
+appVersion: "2.1.0"
diff --git a/helm-charts/common/tgi/gaudi-values.yaml b/helm-charts/common/tgi/gaudi-values.yaml
@@ -13,7 +13,7 @@ image:
   repository: ghcr.io/huggingface/tgi-gaudi
   pullPolicy: IfNotPresent
   # Overrides the image tag whose default is the chart appVersion.
-  tag: "1.2.1"
+  tag: "2.0.1"
 
 imagePullSecrets: []
 nameOverride: ""
@@ -48,8 +48,7 @@ tolerations: []
 
 affinity: {}
 
-LLM_MODEL_ID: ise-uiuc/Magicoder-S-DS-6.7B
-# LLM_MODEL_ID: /data/Magicoder-S-DS-6.7B
+LLM_MODEL_ID: Intel/neural-chat-7b-v3-3
 
 global:
   http_proxy: ""

diff --git a/helm-charts/common/tgi/values.yaml b/helm-charts/common/tgi/values.yaml
@@ -13,7 +13,7 @@ image:
   repository: ghcr.io/huggingface/text-generation-inference
   pullPolicy: IfNotPresent
   # Overrides the image tag whose default is the chart appVersion.
-  tag: "1.4"
+  tag: "2.1.0"
 
 imagePullSecrets: []
 nameOverride: ""
@@ -56,8 +56,7 @@ tolerations: []
 
 affinity: {}
 
-LLM_MODEL_ID: bigscience/bloom-560m
-# LLM_MODEL_ID: /data/OpenCodeInterpreter-DS-6.7B
+LLM_MODEL_ID: Intel/neural-chat-7b-v3-3
 
 global:
   http_proxy: ""

diff --git a/manifests/common/tgi.yaml b/manifests/common/tgi.yaml
@@ -11,10 +11,10 @@ metadata:
     helm.sh/chart: tgi-0.8.0
     app.kubernetes.io/name: tgi
     app.kubernetes.io/instance: tgi
-    app.kubernetes.io/version: "1.4"
+    app.kubernetes.io/version: "2.1.0"
     app.kubernetes.io/managed-by: Helm
 data:
-  MODEL_ID: "bigscience/bloom-560m"
+  MODEL_ID: "Intel/neural-chat-7b-v3-3"
   PORT: "2080"
   HUGGING_FACE_HUB_TOKEN: "insert-your-huggingface-token-here"
   HF_TOKEN: "insert-your-huggingface-token-here"
@@ -40,7 +40,7 @@ metadata:
     helm.sh/chart: tgi-0.8.0
     app.kubernetes.io/name: tgi
     app.kubernetes.io/instance: tgi
-    app.kubernetes.io/version: "1.4"
+    app.kubernetes.io/version: "2.1.0"
     app.kubernetes.io/managed-by: Helm
 spec:
   type: ClusterIP
@@ -65,7 +65,7 @@ metadata:
     helm.sh/chart: tgi-0.8.0
     app.kubernetes.io/name: tgi
     app.kubernetes.io/instance: tgi
-    app.kubernetes.io/version: "1.4"
+    app.kubernetes.io/version: "2.1.0"
     app.kubernetes.io/managed-by: Helm
 spec:
   replicas: 1
@@ -91,7 +91,7 @@ spec:
                 optional: true
           securityContext:
             {}
-          image: "ghcr.io/huggingface/text-generation-inference:1.4"
+          image: "ghcr.io/huggingface/text-generation-inference:2.1.0"
           imagePullPolicy: IfNotPresent
           volumeMounts:
             - mountPath: /data

diff --git a/manifests/common/tgi_gaudi.yaml b/manifests/common/tgi_gaudi.yaml
@@ -11,10 +11,10 @@ metadata:
     helm.sh/chart: tgi-0.8.0
     app.kubernetes.io/name: tgi
     app.kubernetes.io/instance: tgi
-    app.kubernetes.io/version: "1.4"
+    app.kubernetes.io/version: "2.1.0"
     app.kubernetes.io/managed-by: Helm
 data:
-  MODEL_ID: "ise-uiuc/Magicoder-S-DS-6.7B"
+  MODEL_ID: "Intel/neural-chat-7b-v3-3"
   PORT: "2080"
   HUGGING_FACE_HUB_TOKEN: "insert-your-huggingface-token-here"
   HF_TOKEN: "insert-your-huggingface-token-here"
@@ -40,7 +40,7 @@ metadata:
     helm.sh/chart: tgi-0.8.0
     app.kubernetes.io/name: tgi
     app.kubernetes.io/instance: tgi
-    app.kubernetes.io/version: "1.4"
+    app.kubernetes.io/version: "2.1.0"
     app.kubernetes.io/managed-by: Helm
 spec:
   type: ClusterIP
@@ -65,7 +65,7 @@ metadata:
     helm.sh/chart: tgi-0.8.0
     app.kubernetes.io/name: tgi
     app.kubernetes.io/instance: tgi
-    app.kubernetes.io/version: "1.4"
+    app.kubernetes.io/version: "2.1.0"
     app.kubernetes.io/managed-by: Helm
 spec:
   replicas: 1
@@ -91,7 +91,7 @@ spec:
                 optional: true
           securityContext:
             {}
-          image: "ghcr.io/huggingface/tgi-gaudi:1.2.1"
+          image: "ghcr.io/huggingface/tgi-gaudi:2.0.1"
           imagePullPolicy: IfNotPresent
           volumeMounts:
             - mountPath: /data