You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: README.md
+7-7Lines changed: 7 additions & 7 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -47,7 +47,7 @@ Follow these instructions to deploy this example to your Azure subscription, try
47
47
48
48
- The subscription selected must have the following quota available in the location you'll select to deploy this implementation.
49
49
50
-
- Azure OpenAI: Standard, GPT-35-Turbo, 25K TPM
50
+
- Azure OpenAI: Standard, GPT-4o-mini, 10K TPM
51
51
- Storage Accounts: 1
52
52
- Total Cluster Dedicated Regional vCPUs: 4
53
53
- Standard DASv4 Family Cluster Dedicated vCPUs: 4
@@ -130,14 +130,14 @@ To test this architecture, you'll be deploying a pre-built Prompt flow. The Prom
130
130
131
131
1. Connect the `extract_query_from_question` Prompt flow step to your Azure OpenAI model deployment.
132
132
133
-
- For **Connection**, select 'aoai' from the dropdown menu. This is your deployed Azure OpenAI instance.
134
-
- For **deployment_name**, select 'gpt35' from the dropdown menu. This is the model you've deployed in that Azure OpenAI instance.
133
+
- For **Connection**, select 'azureaiservices_aoai' from the dropdown menu. This is your deployed Azure OpenAI instance.
134
+
- For **deployment_name**, select 'gpt4o' from the dropdown menu. This is the model you've deployed in that Azure OpenAI instance.
135
135
- For **response_format**, select '{"type":"text"}' from the dropdown menu
136
136
137
137
1. Connect the `augmented_chat` Prompt flow step to your Azure OpenAI model deployment.
138
138
139
-
- For **Connection**, select the same 'aoai' from the dropdown menu.
140
-
- For **deployment_name**, select the same 'gpt35' from the dropdown menu.
139
+
- For **Connection**, select the same 'azureaiservices_aoai' from the dropdown menu.
140
+
- For **deployment_name**, select the same 'gpt4o' from the dropdown menu.
141
141
- For **response_format**, also select '{"type":"text"}' from the dropdown menu.
142
142
143
143
1. Click **Save** on the flow.
@@ -186,7 +186,7 @@ Here you'll take your tested flow and deploy it to a managed online endpoint.
186
186
There is a notice on the final screen that says:
187
187
188
188
> Following connection(s) are using Microsoft Entra ID based authentication. You need to manually grant the endpoint identity access to the related resource of these connection(s).
189
-
> -aoai
189
+
> -azureaiservices_aoai
190
190
191
191
This has already been taken care of by your IaC deployment. The managed online endpoint identity already has this permission to Azure OpenAI, so there is no action for you to take.
192
192
@@ -240,7 +240,7 @@ az group delete -n $RESOURCE_GROUP -y
240
240
241
241
# Purge the soft delete resources
242
242
az keyvault purge -n kv-${BASE_NAME} -l $LOCATION
243
-
az cognitiveservices account purge -g $RESOURCE_GROUP -l $LOCATION -n oai-${BASE_NAME}
243
+
az cognitiveservices account purge -g $RESOURCE_GROUP -l $LOCATION -n ais-${BASE_NAME}
@description('Assign your user the ability to invoke models in Azure OpenAI. This is needed to execute the Prompt flow from within in the Azure AI Foundry portal.')
description: 'Hub to support the Microsoft Learn Azure OpenAI basic chat implementation. https://learn.microsoft.com/azure/architecture/ai-ml/architecture/basic-openai-e2e-chat'
161
161
publicNetworkAccess: 'Enabled'// Production readiness change: The "Baseline" architecture adds ingress and egress network control over this "Basic" implementation.
162
162
ipAllowlist: []
163
+
networkAcls: {
164
+
defaultAction: 'Allow'// Production readiness change: The "Baseline" architecture adds ingress and egress network control over this "Basic" implementation.
165
+
ipRules: []
166
+
}
163
167
serverlessComputeSettings: null
164
168
enableServiceSideCMKEncryption: false
169
+
provisionNetworkNow: false
165
170
managedNetwork: {
166
171
isolationMode: 'Disabled'// Production readiness change: The "Baseline" architecture adds ingress and egress network control over this "Basic" implementation.
peRequirement: 'NotRequired'// Production readiness change: The "Baseline" architecture adds ingress and egress network control over this "Basic" implementation.
description: 'Project to contain the "Chat with Wikipedia" example Prompt flow that is used as part of the Microsoft Learn Azure OpenAI basic chat implementation. https://learn.microsoft.com/azure/architecture/ai-ml/architecture/basic-openai-e2e-chat'
240
245
v1LegacyMode: false
241
-
publicNetworkAccess: 'Enabled'
246
+
hbiWorkspace: false
247
+
allowRoleAssignmentOnRG: false// Require role assignments at the resource level.
248
+
enableDataIsolation: true
249
+
systemDatastoresAuthMode: 'identity'
250
+
enableServiceSideCMKEncryption: false
251
+
publicNetworkAccess: 'Enabled'// Production readiness change: The "Baseline" architecture adds ingress and egress network control over this "Basic" implementation.
@description('Assign the online endpoint the ability to invoke models in Azure OpenAI. This is needed to execute the Prompt flow from the managed endpoint.')
publicNetworkAccess: 'Enabled'// Production readiness change: This sample uses identity as the perimeter. Production scenarios should layer in network perimeter control as well.
versionUpgradeOption: 'OnceNewDefaultVersionAvailable'// Production readiness change: Always be explicit about model versions, use 'NoAutoUpgrade' to prevent version changes.
0 commit comments