add deploy endpoint workflow

Author: pirocheto

.github/workflows/create_model_endpoint.yaml

@@ -0,0 +1,52 @@
+name: Create Sagemeker Model Endpoint
+
+on:
+  push:
+    branches:
+      - test
+      - main
+  workflow_dispatch:
+
+permissions:
+  id-token: write
+  contents: read
+
+env:
+  AWS_REGION: us-east-1
+  ENVIRONMENT: ${{ github.ref_name == 'main' && 'prod' || github.ref_name == 'test' && 'test' }}
+
+jobs:
+  create_model_endpoint:
+    runs-on: ubuntu-latest
+
+    defaults:
+      run:
+        working-directory: terraform
+
+    steps:
+      - uses: actions/checkout@v3
+
+      - name: Configure AWS Credentials
+        uses: aws-actions/configure-aws-credentials@v4
+        with:
+          audience: sts.amazonaws.com
+          aws-region: ${{ env.AWS_REGION }}
+          role-to-assume: arn:aws:iam::639269844451:role/github-actions
+          role-session-name: ${{ env.ENVIRONMENT }}-deployment
+
+      - name: Setup Terraform
+        uses: hashicorp/setup-terraform@v3
+
+      - name: Terraform Init
+        run: terraform init
+
+      - name: Select/Create Terraform Workspace
+        run: |
+          terraform workspace list || echo "Workspace list unavailable"
+          terraform workspace select --or-create ${{ env.ENVIRONMENT }}
+
+      - name: Terraform Plan
+        run: terraform plan
+
+      - name: Terraform Apply
+        run: terraform apply -auto-approve

.github/workflows/model-to-s3-from-hf-hub.yaml renamed to .github/workflows/push_model_to_s3.yaml

@@ -5,9 +5,7 @@ locals {
 module "model" {
   source               = "./modules/aws-sagemaker-hf-pytorch-inference-model-deployment"
   aws_region           = "us-east-1"
-  model_author_name    = "tabularisai"
   model_name           = "multilingual-sentiment-analysis"
-  model_src            = abspath("../model.tar.gz")
   python_version       = "py310"
   ubuntu_version       = "22.04"
   transformers_version = "4.37.0"

terraform/main.tf

@@ -8,13 +8,6 @@ locals {
   ecr_image_tag  = "${var.pytorch_version}-transformers${var.transformers_version}-cpu-${var.python_version}-ubuntu${var.ubuntu_version}"
 }
 
-resource "aws_s3_object" "model_upload" {
-  bucket = var.model_bucket_name
-  key    = var.model_bucket_key
-  source = var.model_src
-  etag   = filemd5(var.model_src)
-}
-
 resource "aws_iam_role" "model_execution_role" {
   name = "${local.env}-sagemaker-${var.model_name}-execution-role"
 
@@ -92,17 +85,9 @@ resource "aws_iam_role_policy_attachment" "ecr_policy_attachment" {
 resource "aws_sagemaker_model" "model" {
   name               = var.model_name
   execution_role_arn = aws_iam_role.model_execution_role.arn
-  # primary_container {
-  #   image          = "${local.ecr_image_name}:${local.ecr_image_tag}"
-  #   model_data_url = "s3://${aws_s3_object.model_upload.bucket}/${aws_s3_object.model_upload.key}"
-  # }
-
   primary_container {
-    image = "763104351884.dkr.ecr.us-east-1.amazonaws.com/huggingface-pytorch-inference:2.1.0-transformers4.37.0-cpu-py310-ubuntu22.04"
-    environment = {
-      HF_MODEL_ID = "distilbert-base-uncased"
-      HF_TASK     = "text-classification"
-    }
+    image          = "${local.ecr_image_name}:${local.ecr_image_tag}"
+    model_data_url = "s3://${var.model_bucket_name}/${var.model_bucket_key}"
   }
 }
 

terraform/modules/aws-sagemaker-hf-pytorch-inference-model-deployment/main.tf

@@ -5,18 +5,10 @@ variable "aws_region" {
   default     = "us-east-1"
 }
 
-variable "model_author_name" {
-  description = "Author name of the model"
-}
-
 variable "model_name" {
   description = "Name of the model"
 }
 
-variable "model_src" {
-  description = "Path to the model.tar.gz file"
-}
-
 variable "python_version" {
   description = "Python version"
 }