Merge pull request opendatahub-io#332 from vaibhavjainwiz/sync_1

Sync release to main

demo/kserve/custom-manifests/caikit/caikit-standalone/caikit-standalone-isvc-grpc.yaml

@@ -10,8 +10,7 @@ spec:
   predictor:
     serviceAccountName: sa
     model:
-      # https://github.com/kserve/modelmesh-serving/blob/main/docs/predictors/setup-storage.md#3-add-a-storage-entry-to-the-storage-config-secret
       modelFormat:
         name: caikit
       runtime: caikit-standalone-runtime-grpc
-      storageUri: s3://modelmesh-example-models/llm/models/flan-t5-small-caikit # single model here: target directory must contain a config.yml
+      storageUri: s3://modelmesh-example-models/models/hf/

demo/kserve/custom-manifests/caikit/caikit-standalone/caikit-standalone-isvc.yaml

@@ -10,8 +10,7 @@ spec:
   predictor:
     serviceAccountName: sa
     model:
-      # https://github.com/kserve/modelmesh-serving/blob/main/docs/predictors/setup-storage.md#3-add-a-storage-entry-to-the-storage-config-secret
       modelFormat:
         name: caikit
       runtime: caikit-standalone-runtime
-      storageUri: s3://modelmesh-example-models/llm/models/flan-t5-small-caikit # single model here: target directory must contain a config.yml
+      storageUri: s3://modelmesh-example-models/models/hf/

demo/kserve/custom-manifests/caikit/caikit-standalone/caikit-standalone-servingruntime-grpc.yaml

@@ -15,8 +15,8 @@ spec:
       env:
         - name: RUNTIME_LOCAL_MODELS_DIR
           value: /mnt/models
-        - name: TRANSFORMERS_CACHE
-          value: /tmp/transformers_cache
+        - name: HF_HOME
+          value: /tmp/hf_home
         - name: RUNTIME_GRPC_ENABLED
           value: "true"
         - name: RUNTIME_HTTP_ENABLED

demo/kserve/custom-manifests/caikit/caikit-standalone/caikit-standalone-servingruntime.yaml

@@ -10,13 +10,13 @@ spec:
       name: caikit
   containers:
     - name: kserve-container
-      image: quay.io/opendatahub/caikit-nlp:stable
+      image: quay.io/opendatahub/caikit-nlp:fast
       command: ["python", "-m", "caikit.runtime"]
       env:
         - name: RUNTIME_LOCAL_MODELS_DIR
           value: /mnt/models
-        - name: TRANSFORMERS_CACHE
-          value: /tmp/transformers_cache
+        - name: HF_HOME
+          value: /tmp/hf_home
         - name: RUNTIME_GRPC_ENABLED
           value: "false"
         - name: RUNTIME_HTTP_ENABLED

demo/kserve/deploy-embeddings-scripts.md

@@ -0,0 +1,57 @@
+## Using scripts to deploy an embeddings model with the Caikit Standalone Serving Runtime
+You can deploy and test a embedding model by running the scripts provided in the the   `caikit-tgis-serving` repo. These scripts deploy a all-MiniLM-L12-v2 model with the Caikit Standalone Serving runtime. This model has already been containerized into an S3 MinIO bucket.
+
+**Prerequisites**
+- You installed the Caikit-TGIS-Serving image as described in the Caikit-TGIS-Serving README file.
+
+- You installed the scripts as described in Script-based Installation.
+
+- Your current working directory is the /caikit-tgis-serving/demo/kserve/ directory.
+
+**Procedure**
+1. Deploy a sample embeddings model
+
+   Replace the default value of `image` in `custom-manifests/minio/minio.yaml` with the contanerized embeddings model.
+
+   ```
+   .
+   .
+   .
+        image: quay.io/christinaexyou/modelmesh-minio-examples:embedding-models
+   ```
+
+   Replace all instances of `tgis` with `standalone` in `scripts/test/deploy-model.sh`
+
+   ```
+   sed 's/tgis/standalone/g' ./scripts/test/deploy-model.sh | tee ./scripts/test/deploy-model-standalone.sh
+   ```
+
+   For HTTP:
+
+   ```
+   ./scripts/test/deploy-model-standalone.sh
+   ```
+
+   For grPC:
+
+   ```
+   ./scripts/test/deploy-model-standalone.sh grpc
+   ```
+2. Perform inference
+
+   For HTTP:
+
+   ```
+   ./scripts/test/http-call-embeddings.sh
+   ```
+
+   For grPC:
+   ```
+   ./scripts/test/grpc-call-embeddings.sh
+   ```
+
+4. Delete the sample model
+
+    ```
+    tests/scripts/delete-model.sh
+    ```