A week ago, AWS announced that Anthropic’s Claude 3 Sonnet model now available on Amazon Bedrock. I am so eager to give it a try, especially its vision capabilities, as it is the first multimodal foundation model in Amazon Bedrock excluding embedding models.
According to Anthropic’s introduction, the Claude 3 family is smarter, faster and safer. The most intelligent model (Claude 3 Opus) outperforms GPT-4.
Regarding the vision capabilities
The Claude 3 models have sophisticated vision capabilities on par with other leading models. They can process a wide range of visual formats, including photos, charts, graphs and technical diagrams. We’re particularly excited to provide this new modality to our enterprise customers, some of whom have up to 50% of their knowledge bases encoded in various formats such as PDFs, flowcharts, or presentation slides.
Sonnet is the only available Claude 3 model in Amazon Bedrock as the time of writing. It strikes the ideal balance between intelligence and speed – particularly for enterprise workloads. I did following tests with Claude 3 Sonnet with Image-Reader which is the project that I created to explore the capabilities of Claude 3. You can not only use it to read images, but also can test prompts (e.g from prompt library) without images.
- Read a product catalog
- Read a housing price diagram
- Read a floor plan
- Read a architecture diagram
- Read and compare two pictures
- Read a pictures of books
References:
Jackie Chen's IT Workshop 
One thought on “Image-Reader: A project to explore Claude 3 Vision Capabilities”