Find answers from the community

Home
Members
mandelbrot4669
m
mandelbrot4669
Offline, last seen 3 months ago
Joined September 25, 2024
Hi everyone, I have been running into a somewhat trivial issue while using GPT4-vision to extract the details from an architectural blueprint jpg. The image itself is in high resolution, and all values are easily visible if you zoom in, however, as the image itself is quite detailed, the values are not legible when zoomed out. When i used GPT-vision-preview through the api on the image, the GPT complained about the values being too small to read. What do I do? This was the answer given by GPT
Plain Text
'message': {'role': 'assistant', 'content': 'You\'ve provided several images of what appear to be architectural plans and elevations for a building. Due to the detail and scale of the plans, specific values and text may not be entirely legible or may be too small to discern accurately.
2 comments
L
m