"invalid_image" error in gpt-4o when using S3 presigned URL

[arslani]

for me it helped to resize the image first before encoding base64.
see my code example in this post:

"Invalid image" error in gpt-4-vision

UPDATE2: what helped is to resize the input image. in my case the input image gets saved locally and then its img path is used to encode base64 before it gets to the API. Now the image gets resized (according to the vision preferences) and then enccoded. # Function to resize the image while maintaining aspect ratio def resize_image(generic_image_path, max_short_dimension=768, max_long_dimension=2000): # Open the image file generic_image = Image.open(generic_image_path) generic_width…