Implementing model fallbacks
In production AI applications, relying on a single model can lead to service interruptions. This exercise implements a fallback from Claude to Nova Text Lite when encountering issues. This implementation ensures continuous service while handling model-specific requirements.
The json
and boto3
libraries, and the bedrock
client, are preloaded.
This exercise is part of the course
Introduction to Amazon Bedrock
Exercise instructions
- Wrap the primary Claude model call in a
try
block. - In the
except
block, handle errors usingClientError
and call the Nova Text Lite model.
Hands-on interactive exercise
Have a go at this exercise by completing this sample code.
def process_with_fallback(prompt):
# Handle errors with a try-except block
____:
response = bedrock.invoke_model(
modelId="anthropic.claude-3-5-sonnet-20240620-v1:0",
body=json.dumps({"anthropic_version": "bedrock-2023-05-31", "max_tokens": 100,
"messages": [{"role": "user", "content": [{"type": "text", "text": prompt}]}]}))
return json.loads(response["body"].read().decode())["content"][0]["text"]
____:
fallback = bedrock.invoke_model(
modelId="amazon.nova-lite-v1:0",
body=json.dumps({"messages": [{"role": "user", "content": [{"text": prompt}]}]}))
return json.loads(fallback["body"].read().decode())["output"]["message"]["content"][0]["text"]
print(process_with_fallback("Explain AWS Lambda"))