Show HN: GPT-4 vision utilities to browse the web

asim-shrestha wrote: post #3

ilaksh wrote: post #2

The README seems mismatched.

Mind elaborating here? Happy to update the README if there are issues

It says Google OCR in the usage example, but your description above mentions GPT-4 vision. So this makes no sense. I see nothing about Open AI API key in the example.

Re: Show HN: GPT-4 vision utilities to browse the web

#5

Post by KhoomeiK » Sat, Nov 11, 2023, 11:20 PM UTC

ilaksh wrote: post #4

asim-shrestha wrote: post #3

Earlier quoted context omitted.

Mind elaborating here? Happy to update the README if there are issues

It says Google OCR in the usage example, but your description above mentions GPT-4 vision. So this makes no sense. I see nothing about Open AI API key in the example.

We use Google OCR to convert the screenshot into whitespace-structured text for unimodal LLMs. For multimodal LLMs like GPT-4V, you don't need to use the OCR utilities. Thanks for the feedback though, we'll clarify!