Update aml deployment

aml/README.md

@@ -2,6 +2,7 @@
 docker build . -t db4c2190dd824d1f950f5d1555fbadf0.azurecr.io/text-generation:0.1
 docker push db4c2190dd824d1f950f5d1555fbadf0.azurecr.io/text-generation:0.1
 
+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
 ```
\ No newline at end of file

aml/deployment.yaml

 $schema: https://azuremlschemas.azureedge.net/latest/managedOnlineDeployment.schema.json
 name: bloom-deployment
 endpoint_name: bloom-inference
-model:
-  name: bloom
-  path: ./bloom
+model: azureml:bloom:1
 model_mount_path: /var/azureml-model
 environment_variables:
   MODEL_BASE_PATH: /var/azureml-model/bloom
@@ -24,6 +22,7 @@ environment:
 instance_type: Standard_ND96amsr_A100_v4
 request_settings:
   request_timeout_ms: 90000
+  max_concurrent_requests_per_instance: 256
 liveness_probe:
   initial_delay: 300
   timeout: 20

aml/endpoint.yaml

 $schema: https://azuremlsdk2.blob.core.windows.net/latest/managedOnlineEndpoint.schema.json
 name: bloom-inference
-auth_mode: aml_token
+auth_mode: key

aml/model.yaml

+$schema: https://azuremlschemas.azureedge.net/latest/model.schema.json
+name: bloom
+version: 1
+path: ./bloom
+type: custom_model