I have a question regarding the AnyRes feature in LLaVA-NeXT-Video. The documentation mentions that AnyRes enables high-resolution image processing. However, when examining the demo code at https://github.com/LLaVA-VL/LLaVA NeXT/blob/main/playground/demo/video_demo.py, I noticed that neither image_grid_pinpoints nor AnyRes are being used. This is in contrast to LLaVA-NeXT-Image, where these features are utilized.
I have a question regarding the AnyRes feature in LLaVA-NeXT-Video. The documentation mentions that AnyRes enables high-resolution image processing. However, when examining the demo code at https://github.com/LLaVA-VL/LLaVA NeXT/blob/main/playground/demo/video_demo.py, I noticed that neither
image_grid_pinpoints
nor AnyRes are being used. This is in contrast to LLaVA-NeXT-Image, where these features are utilized.