No post body was provided.
Untitled topic
1–5 of 5 posts
Re: undefined
#2Transformers seem great. IMO the big challenge is still to better understand how they perform in operation.
Re: undefined
#3The fundamental architectures of transformers haven't changed in years. I wonder what kind of design changes will be made in the future and how that impacts these results.
Re: undefined
#4Interesting, thanks for sharing! Somehow I'm not surprised. My experience building systems for real world applications is that choosing a simple CNN is usually the way to go, and it's all in the data. Choosing more complex model classes also comes at a cost, higher variance and more likelihood of issues related to robustness and generalization.
Re: undefined
#5Interesting, but what about the more recent YOLO versions? Do you have any research on if they encounter the same issues?