update cfn

Author: aws-banjo

ai-gen-workshop-cfn.yml

@@ -1,22 +1,17 @@
-# Copyright 2023 Amazon.com and its affiliates; all rights reserved. 
+# Copyright 2023 Amazon.com and its affiliates; all rights reserved.
 # This file is Amazon Web Services Content and may not be duplicated or distributed without permission.
 AWSTemplateFormatVersion: "2010-09-09"
 Description: SageMaker Foundational Models Workshop
 
-
 Parameters:
   UserProfileName:
     Type: String
     Description: The user profile name for the SageMaker workshop
-    Default: 'workshop-builder'
+    Default: "workshop-builder"
   DomainName:
     Type: String
-    Description: The domain name of the Sagemaker studio instance
-    Default: 'd-aigensagemaker'
-  DockerImage:
-    Type: String
-    Description: TODO you must build the image, in the repo and push to your ECR -https://docs.aws.amazon.com/AmazonECR/latest/userguide/docker-push-ecr-image.html
-    Default: 'TODO'
+    Description: The domain name of the SageMaker studio instance
+    Default: "d-aigensagemaker"
 
 Mappings:
   RegionMap:
@@ -78,7 +73,7 @@ Mappings:
       datascience: "arn:aws:sagemaker:sa-east-1:782484402741:image/datascience-1.0"
       datawrangler: "arn:aws:sagemaker:sa-east-1:424196993095:image/sagemaker-data-wrangler-1.0"
 
-Resources: 
+Resources:
   LambdaExecutionRole:
     Type: "AWS::IAM::Role"
     Properties:
@@ -93,11 +88,11 @@ Resources:
               - "sts:AssumeRole"
       Path: /
       ManagedPolicyArns:
-        - 'arn:aws:iam::aws:policy/AmazonEC2ReadOnlyAccess'
-        - 'arn:aws:iam::aws:policy/AmazonS3FullAccess'
-        - 'arn:aws:iam::aws:policy/IAMFullAccess'
-        - 'arn:aws:iam::aws:policy/service-role/AWSLambdaBasicExecutionRole'
-        - 'arn:aws:iam::aws:policy/AmazonSageMakerFullAccess'
+        - "arn:aws:iam::aws:policy/AmazonEC2ReadOnlyAccess"
+        - "arn:aws:iam::aws:policy/AmazonS3FullAccess"
+        - "arn:aws:iam::aws:policy/IAMFullAccess"
+        - "arn:aws:iam::aws:policy/service-role/AWSLambdaBasicExecutionRole"
+        - "arn:aws:iam::aws:policy/AmazonSageMakerFullAccess"
 
   SageMakerExecutionRole:
     Type: AWS::IAM::Role
@@ -107,8 +102,7 @@ Resources:
           PolicyDocument:
             Version: 2012-10-17
             Statement:
-              -
-                Effect: Allow
+              - Effect: Allow
                 Action:
                   - s3:GetObject
                   - s3:PutObject
@@ -119,42 +113,40 @@ Resources:
           PolicyDocument:
             Version: 2012-10-17
             Statement:
-              -
-                Effect: Allow
+              - Effect: Allow
                 Action:
                   - iam:GetRole
                   - iam:GetRolePolicy
-                Resource: '*'
+                Resource: "*"
       AssumeRolePolicyDocument:
         Version: 2012-10-17
         Statement:
-          -
-            Effect: Allow
+          - Effect: Allow
             Principal:
               Service:
                 - sagemaker.amazonaws.com
             Action:
               - sts:AssumeRole
       ManagedPolicyArns:
-        - 'arn:aws:iam::aws:policy/AmazonSageMakerFullAccess'
-        - 'arn:aws:iam::aws:policy/AmazonS3FullAccess'
-        - 'arn:aws:iam::aws:policy/AWSCloudFormationFullAccess'
-        - 'arn:aws:iam::aws:policy/AWSCodePipeline_FullAccess'
+        - "arn:aws:iam::aws:policy/AmazonSageMakerFullAccess"
+        - "arn:aws:iam::aws:policy/AmazonS3FullAccess"
+        - "arn:aws:iam::aws:policy/AWSCloudFormationFullAccess"
+        - "arn:aws:iam::aws:policy/AWSCodePipeline_FullAccess"
 
   DomainDeploymentDelay:
-    Type: 'Custom::Delay'
+    Type: "Custom::Delay"
     DependsOn: SageMakerExecutionRole
     Properties:
       ServiceToken: !GetAtt DelayLambda.Arn
       TimeToWait: 300
 
   DelayLambda:
-    Type: 'AWS::Lambda::Function'
+    Type: "AWS::Lambda::Function"
     Properties:
       Handler: "index.handler"
       Timeout: 660
       Role: !GetAtt LambdaExecutionRole.Arn
-      Runtime: python3.7
+      Runtime: python3.12
       Code:
         ZipFile: |
           import json
@@ -211,14 +203,13 @@ Resources:
       Handler: index.lambda_handler
       MemorySize: 512
       Role: !GetAtt LambdaExecutionRole.Arn
-      Runtime: python3.7
+      Runtime: python3.12
       Timeout: 5
   DefaultVpcFinder:
     Type: Custom::ResourceForFindingDefaultVpc
     Properties:
       ServiceToken: !GetAtt DefaultVpcLambda.Arn
 
-
   FmCreateImage:
     Type: AWS::Lambda::Function
     Properties:
@@ -228,29 +219,19 @@ Resources:
           import json
           import boto3
 
-          def parse_response(query_response):
-              """Parse response and return generated image and the prompt"""
-
-              response_dict = json.loads(query_response)
-              return response_dict["generated_image"], response_dict["prompt"]
-
-
           def lambda_handler(event, context):
 
               client = boto3.client('sagemaker-runtime')
               
               data_string = event["body"]
               text = json.loads(data_string)["text"]
-              
-              encoded_text = text.encode("utf-8")
-              
-              
+                  
               CONTENT_TYPE = 'application/x-text'
-              endpoint_name = "YOUR_MODEL_ENDPOINT" # Replace with your model endpoint
-              
-              response = client.invoke_endpoint(EndpointName=endpoint_name, 
-                                            ContentType=CONTENT_TYPE, 
-                                            Body=encoded_text)
+              endpoint_name = "REPLACE_WITH_MODEL_ENDPOINT" # Replace with your model endpoint
+
+              encoded_text = text.encode("utf-8")
+              response = client.invoke_endpoint(EndpointName=endpoint_name, InferenceComponentName='model-txt2img-stabilityai-stable-diffusion-v2-1-20250203-192659', ContentType='application/x-text', Body=encoded_text, Accept='application/json')
+
 
               print(response)
 
@@ -267,7 +248,7 @@ Resources:
       Handler: index.lambda_handler
       MemorySize: 512
       Role: !GetAtt LambdaExecutionRole.Arn
-      Runtime: python3.9
+      Runtime: python3.12
       Timeout: 200
 
   FmCreateSummary:
@@ -286,11 +267,11 @@ Resources:
               data_string = event["body"]
               text = json.loads(data_string)["text"]
 
-              query = 'write a summary'
-              
-              prompt = f'{text}\n{query}'
+              query = '<|im_start|>system\nyou are an expert summary writter<|im_end|>\n<|im_start|>user\nwrite a summary of the following text:'
+
+              prompt = f'{query}\n{text}<|im_end|>\n<|im_start|>assistant:'
               
-              MAX_LENGTH = 150 # 256
+              MAX_LENGTH = 1000
               NUM_RETURN_SEQUENCES = 1
               TOP_K = 0
               TOP_P = 0.7
@@ -308,39 +289,39 @@ Resources:
 
               payload = json.dumps(payload).encode('utf-8')
               
-              endpoint_name = "YOUR_MODEL_ENDPOINT" # Replace with your model endpoint
+              endpoint_name = "REPLACE_WITH_MODEL_ENDPOINT" # Replace with your model endpoint
               
               response = client.invoke_endpoint(EndpointName=endpoint_name, 
                                             ContentType=CONTENT_TYPE, 
                                             Body=payload)
               model_predictions = json.loads(response['Body'].read())
               generated_text = model_predictions[0]["generated_text"]
               print(f'Response: {generated_text}')
+              final_response = generated_text.split("assistant:")[1].strip()
               
-              resp = json.dumps({"text":generated_text})
+              resp = json.dumps({"text":final_response})
               
               
               return {
                   'statusCode': 200,
                   'body': resp
               }
+
       Description: Call model to generate summary
       Handler: index.lambda_handler
       MemorySize: 512
       Role: !GetAtt LambdaExecutionRole.Arn
-      Runtime: python3.9
+      Runtime: python3.12
       Timeout: 200
 
-
-
   StudioDomain:
     Type: AWS::SageMaker::Domain
     DependsOn: DomainDeploymentDelay
     Properties:
       AppNetworkAccessType: PublicInternetOnly
       AuthMode: IAM
       DefaultUserSettings:
-          ExecutionRole: !GetAtt SageMakerExecutionRole.Arn
+        ExecutionRole: !GetAtt SageMakerExecutionRole.Arn
       DomainName: !Ref DomainName
       SubnetIds: !GetAtt DefaultVpcFinder.Subnets
       VpcId: !GetAtt DefaultVpcFinder.VpcId
@@ -362,21 +343,6 @@ Resources:
       DomainId: !GetAtt StudioDomain.DomainId
       UserProfileName: !Ref UserProfileName
 
-  DataScienceApp:
-    Type: AWS::SageMaker::App
-    DependsOn: UserProfile
-    Properties:
-      AppName: instance-event-engine-datascience-ml-t3-medium
-      AppType: KernelGateway
-      DomainId: !GetAtt StudioDomain.DomainId
-      ResourceSpec:
-        InstanceType:  ml.t3.medium
-        SageMakerImageArn: !FindInMap
-          - RegionMap
-          - !Ref 'AWS::Region'
-          - datascience
-      UserProfileName: !Ref UserProfileName
-
   ContainerCluster:
     Type: AWS::ECS::Cluster
     Properties:
@@ -396,11 +362,11 @@ Resources:
           - Effect: Allow
             Principal:
               Service: ecs-tasks.amazonaws.com
-            Action: 'sts:AssumeRole'
+            Action: "sts:AssumeRole"
       ManagedPolicyArns:
-        - 'arn:aws:iam::aws:policy/service-role/AmazonECSTaskExecutionRolePolicy'
-        - 'arn:aws:iam::aws:policy/AmazonElasticContainerRegistryPublicReadOnly'
-        - 'arn:aws:iam::aws:policy/AWSLambda_FullAccess'
+        - "arn:aws:iam::aws:policy/service-role/AmazonECSTaskExecutionRolePolicy"
+        - "arn:aws:iam::aws:policy/AmazonElasticContainerRegistryPublicReadOnly"
+        - "arn:aws:iam::aws:policy/AWSLambda_FullAccess"
 
   TaskRole:
     Type: AWS::IAM::Role
@@ -411,12 +377,11 @@ Resources:
           - Effect: Allow
             Principal:
               Service: ecs-tasks.amazonaws.com
-            Action: 'sts:AssumeRole'
+            Action: "sts:AssumeRole"
       ManagedPolicyArns:
-        - 'arn:aws:iam::aws:policy/AWSGlueSchemaRegistryFullAccess'
-        - 'arn:aws:iam::aws:policy/service-role/AmazonECSTaskExecutionRolePolicy'
-        - 'arn:aws:iam::aws:policy/AWSLambda_FullAccess'
-
+        - "arn:aws:iam::aws:policy/AWSGlueSchemaRegistryFullAccess"
+        - "arn:aws:iam::aws:policy/service-role/AmazonECSTaskExecutionRolePolicy"
+        - "arn:aws:iam::aws:policy/AWSLambda_FullAccess"
 
   BuildOnTaskDefinition:
     Type: AWS::ECS::TaskDefinition
@@ -431,12 +396,12 @@ Resources:
       ExecutionRoleArn: !Ref ExecutionRole
       ContainerDefinitions:
         - Name: build-on-poster
-          Image: !Ref DockerImage
+          Image: public.ecr.aws/o3s1z9h7/build-on-poster:latest
           Environment:
             - Name: IMAGE_API
-              Value: ''
+              Value: ""
             - Name: TEXT_API
-              Value: ''
+              Value: ""
 
           LogConfiguration:
             LogDriver: awslogs