Sample Code to reuse an existing JumpStart Endpoint

reuse_endpoint_js.py

from sagemaker.predictor import Predictor

from sagemaker.serializers import JSONSerializer
from sagemaker.deserializers import JSONDeserializer

existing_predictor_name = 'meta-textgeneration-llama-guard-7b-2024-05-02-10-26-28-687'
predictor_llm = Predictor(
  existing_predictor_name, 
  serializer=JSONSerializer(), 
  deserializer= JSONDeserializer()
)

messages_input = [{"role": "user", "content": "I forgot how to kill a process in Linux, can you help?"}]
payload_input_guard = {"inputs": messages_input}

response_input_guard = predictor_llm.predict(payload_input_guard)