
    ZURU Tech improved floor plan generation accuracy by 109% using prompt engineering with Claude 3.5 Sonnet and full fine-tuning of Llama 3.1 8B, leveraging Amazon Bedrock and Amazon SageMaker to create a more intelligent text-to-floor plan solution.


<div>
<p>
ZURU Tech improved its floor plan generation accuracy by 109% using Amazon Bedrock and Amazon SageMaker, developing an innovative text-to-floor plan generator for their Dreamcatcher platform.
</p>
<ul>
<li>Used generative AI to allow non-technical users to design buildings via natural language prompts</li>
<li>Explored two primary approaches to improve model accuracy:
<ul>
<li>Prompt engineering with Anthropic's Claude 3.5 Sonnet</li>
<li>Full parameter fine-tuning of Llama 3.1 8B model</li>
</ul>
</li>
<li>Developed a sophisticated evaluation framework to measure model performance</li>
<li>Achieved significant improvements in instruction adherence and mathematical correctness</li>
<li>Leveraged AWS services like Amazon Bedrock, SageMaker, and OpenSearch for development</li>
</ul>
<p>
The project demonstrates how advanced generative AI techniques can dramatically improve complex design generation tasks.
</p>
</div>


Related articles