What is the default maximum resolution of the image size read?
#17
by
yang4
- opened
I tested GLM's detection performance on images and found that GLM has strong representation and positioning capabilities, but the output results do not correspond. I guess it is because the image is compressed during the processing. What is the maximum value of image_resolution in GLM? Does it use the same equal-area scaling strategy as Qwen?
Note that I input two pictures that are less than or equal to 2k