sync with main

.gitignore

@@ -1,3 +1,4 @@
 .venvIB*
 .tmp
 .gitignore
+*.pyc

README.md

@@ -71,7 +71,7 @@ InferenceBenchmarker takes your invocation logic defined in Python functions in
         # user on that worker. Put expensive, reusable client state here — e.g. the boto3
         # client and its connection pool — so it isn't rebuilt per request.
         import json, boto3
-        client = boto3.client('sagemaker-runtime', region_name='us-east-1')
+        client = boto3.client('sagemaker-runtime')
 
         def invoke(payload: Payload) -> None:
             # PER-REQUEST: called once for each request a user fires.
@@ -279,7 +279,7 @@ See [client_capacity/README.md](client_capacity/README.md) for usage.
 
 * **Interactive CLI** – Add traces while running benchmarks in the current dry benchmark tool. [tracking issues](https://github.com/aws-samples/sample-InferenceBenchmarker/issues)
 
-* **Automatic RPS** – Automate trial and error server rps supported at success threshold when --endpoint-config for hardware telemetry provided. [tracking issue](https://github.com/aws-samples/sample-InferenceBenchmarker/issues/2)
+* **Automatic RPS** – Automate trial and error server rps supported at success threshold when --endpoint-config for hardware telemetry provided. [tracking issue](https://github.com/aws-samples/sample-InferenceBenchmarker/issues/3)
 
 * **Plot metadata** – Provide a dict to add legend and hover info for better visuals in plots. [tracking issues](https://github.com/aws-samples/sample-InferenceBenchmarker/issues)
 

factories/sagemakerai_realtime/factories_cnn.py

@@ -10,7 +10,6 @@ def invoke_factory(endpoint_name=None):
 
     client = boto3.client(
         'sagemaker-runtime',
-        region_name='us-east-1',
         config=Config(retries={'max_attempts': 3}, max_pool_connections=500, read_timeout=3600),
     )
 

factories/sagemakerai_realtime/factories_ic.py

@@ -13,7 +13,6 @@ def invoke_factory(endpoint_name=None):
 
     client = boto3.client(
         'sagemaker-runtime',
-        region_name='us-east-1',
         config=Config(retries={'max_attempts': 0}, max_pool_connections=500),
     )
 

factories/sagemakerai_realtime/factories_llm_textgeneration.py

@@ -7,7 +7,6 @@ def invoke_factory(endpoint_name=None):
 
     client = boto3.client(
         'sagemaker-runtime',
-        region_name='us-east-1',
         config=Config(retries={'max_attempts': 3}, max_pool_connections=500, read_timeout=3600),
     )