Gemma is now available in Amazon SageMaker JumpStart

This blog post announces the availability of the Gemma language model in Amazon SageMaker JumpStart, and walks through how to deploy and fine-tune it for various use cases.

Specifically, the article covers:

• Background on the Gemma model family and its capabilities
• How to deploy the pre-trained Gemma model for inference using SageMaker JumpStart
• Example payloads and outputs for tasks like code generation and creative writing
• How to fine-tune Gemma on a conversational dataset using QLoRA
• Evaluating and deploying the fine-tuned Gemma model
• Benchmarking latency and throughput for different input sizes
• Conclusion and next steps