3 ms·
It certainly isn't the same cost if expressed as a non-subsidized $$$ one needs for the Transformers compute aka infra. CNNs trained specifically for OCR can r
by menaerus 2y ago
It certainly isn't the same cost if expressed as a non-subsidized $$$ one needs for the Transformers compute aka infra.
CNNs trained specifically for OCR can run in real time on as small compute as a mobile device is.
- anon373839 2y agoA bit of a tangent, but aren’t CNNs still dominating over ViTs among computer vision competition winners?
- menaerus 2y agoI haven't watched that space very closely but IMO ViTs have a great potential to extract from since in comparison to CNNs they allow the model to learn and understand complex relations in the data. Where this matters, I expect it to matter a lot. OCR I think is not the greatest such example - while it matters to understand the surrounding context, I think it's not that critical for performance.