From e114d8748673a067324e0809479273d522a614e1 Mon Sep 17 00:00:00 2001
From: OlivierDehaene <olivier@huggingface.co>
Date: Mon, 6 Feb 2023 14:33:56 +0100
Subject: [PATCH] feat(ci): push to AML registry (#56)

---
 .github/workflows/build.yaml |  7 +++++++
 aml/README.md                | 13 ++++++++++---
 aml/deployment.yaml          |  6 +++---
 aml/model.yaml               |  4 +---
 4 files changed, 21 insertions(+), 9 deletions(-)

diff --git a/.github/workflows/build.yaml b/.github/workflows/build.yaml
index d89f0982..8304c8d1 100644
--- a/.github/workflows/build.yaml
+++ b/.github/workflows/build.yaml
@@ -35,6 +35,12 @@ jobs:
           username: ${{ secrets.TAILSCALE_DOCKER_USERNAME }}
           password: ${{ secrets.TAILSCALE_DOCKER_PASSWORD }}
           registry: registry.internal.huggingface.tech
+      - name: Login to Azure Container Registry
+        uses: docker/login-action@v2.1.0
+        with:
+          username: ${{ secrets.AZURE_DOCKER_USERNAME }}
+          password: ${{ secrets.AZURE_DOCKER_PASSWORD }}
+          registry: db4c2190dd824d1f950f5d1555fbadf0.azurecr.io
       - name: Extract metadata (tags, labels) for Docker
         id: meta
         uses: docker/metadata-action@v4.3.0
@@ -44,6 +50,7 @@ jobs:
           images: |
             ghcr.io/huggingface/text-generation-inference
             registry.internal.huggingface.tech/api-inference/community/text-generation-inference
+            db4c2190dd824d1f950f5d1555fbadf0.azurecr.io/text-generation-inference
           tags: |
             type=semver,pattern={{version}}
             type=semver,pattern={{major}}.{{minor}}
diff --git a/aml/README.md b/aml/README.md
index 959e2942..8e78b0ab 100644
--- a/aml/README.md
+++ b/aml/README.md
@@ -1,8 +1,15 @@
-```shell
-docker build . -t db4c2190dd824d1f950f5d1555fbadf0.azurecr.io/text-generation:0.1
-docker push db4c2190dd824d1f950f5d1555fbadf0.azurecr.io/text-generation:0.1
+# Azure ML endpoint
 
+## Create all resources
+
+```shell
 az ml model create -f model.yaml -g HuggingFace-BLOOM-ModelPage -w HuggingFace
 az ml online-endpoint create -f endpoint.yaml -g HuggingFace-BLOOM-ModelPage -w HuggingFace
 az ml online-deployment create -f deployment.yaml -g HuggingFace-BLOOM-ModelPage -w HuggingFace
+```
+
+## Update deployment
+
+```shell
+az ml online-deployment update -f deployment.yaml -g HuggingFace-BLOOM-ModelPage -w HuggingFace
 ```
\ No newline at end of file
diff --git a/aml/deployment.yaml b/aml/deployment.yaml
index 16ef3dc7..320eba24 100644
--- a/aml/deployment.yaml
+++ b/aml/deployment.yaml
@@ -1,14 +1,14 @@
 $schema: https://azuremlschemas.azureedge.net/latest/managedOnlineDeployment.schema.json
 name: bloom-deployment
 endpoint_name: bloom-inference
-model: azureml:bloom:1
+model: azureml:bloom-safetensors:1
 model_mount_path: /var/azureml-model
 environment_variables:
-  HUGGINGFACE_HUB_CACHE: /var/azureml-model/bloom
+  WEIGHTS_CACHE_OVERRIDE: /var/azureml-model/bloom-safetensors
   MODEL_ID: bigscience/bloom
   NUM_SHARD: 8
 environment:
-  image: db4c2190dd824d1f950f5d1555fbadf0.azurecr.io/text-generation-inference:0.3.1
+  image: db4c2190dd824d1f950f5d1555fbadf0.azurecr.io/text-generation-inference:0.2.0
   inference_config:
     liveness_route:
       port: 80
diff --git a/aml/model.yaml b/aml/model.yaml
index bd490f1a..bfcdd33f 100644
--- a/aml/model.yaml
+++ b/aml/model.yaml
@@ -1,5 +1,3 @@
 $schema: https://azuremlschemas.azureedge.net/latest/model.schema.json
 name: bloom-safetensors
-version: 1
-path: ./bloom-safetensors
-type: custom_model
+path: /data/bloom-safetensors