24 lines
683 B
Python
24 lines
683 B
Python
from mlflow.deployments import get_deploy_client
|
|
|
|
|
|
def main():
|
|
client = get_deploy_client("http://localhost:7000")
|
|
|
|
print(f"Bedrock endpoints: {client.list_endpoints()}\n")
|
|
print(f"Bedrock completions endpoint info: {client.get_endpoint(endpoint='completions')}\n")
|
|
|
|
# Completions example
|
|
response_completions = client.predict(
|
|
endpoint="completions",
|
|
inputs={
|
|
"prompt": "How many patties could be stacked on a cheeseburger before issues arise?",
|
|
"max_tokens": 200,
|
|
"temperature": 0.25,
|
|
},
|
|
)
|
|
print(f"Bedrock completions response: {response_completions}")
|
|
|
|
|
|
if __name__ == "__main__":
|
|
main()
|