[2010.11929] An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale
5
Public whispers
5
Contributors
2026-07-17 10:04:36
First whispered
Public whispers on this page
Text Highlight2026-07-17 13:25:36
Original Highlight Excerpt
"attention is either applied in conjunctio"
Whisper Note
I always thought CNN+attention was the way, guess this flips it.
Text Highlight2026-07-17 13:16:36
Original Highlight Excerpt
"attention is either applied in conjunctio"
Whisper Note
So they're saying attention doesn't need convnets to lean on? That's a bold claim.
Text Highlight2026-07-17 10:22:36
Original Highlight Excerpt
"An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale"
Whisper Note
I tried this once, my GPU cried. Paper's good though.
Text Highlight2026-07-17 10:13:36
Original Highlight Excerpt
"An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale"
Whisper Note
16x16 words, more like 256 patches, but okay, catchy title.
Text Highlight2026-07-17 10:04:36
Original Highlight Excerpt
"An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale"
Whisper Note
So transformers finally came for images, huh? wild times.
Share this page's whispers
Short link
https://domwhisper.com/s/4aba4475c526Embed snippet
<iframe src="https://domwhisper.com/embed/4aba4475c526" width="100%" height="480" style="border:0;border-radius:16px" loading="lazy"></iframe>See what people are discussing on arxiv.org
Install DomWhisper to view live whispers as you browse, and join the discussion.
Get the extension