We’ve all rightfully been taken by LLMs and how impressive the current state-of-the-art is. But, don’t sleep on computer vision. The progress is similarly insane.

We’ve got a breakdown of the changes: A how to train tutorial: And over 50 fine-tuned YOLOv8 models shared by our community that you can try in your browser (on images, video, or webcam):

i tried it a bit last night. it's quite good!

@marianojavierd1 In contrast to LLM there are some pretty objective benchmarks for fi segmentation on coco and I feel there’s still a loooong road ahead.

Terminator vibes

@bobpoekert ignore all the flashing and "the boot"

Does it process all that in real time, or does it slowly go through frame by frame?

boat train bus boat

Next up…