I really appreciate that OpenAI actually did that in the CLIP paper, by reporting the result on a 15M subset of YFCC + releasing the actual subset. It allowed us nice comparisons and sanity-checks in a recent paper, and allows trustable independent reproduction à la OpenCLIP. 👍
I really appreciate that OpenAI actually did that in the CLIP paper, by reporting the result on a 15M subset of YFCC + releasing the actual subset. It allowed us nice comparisons and sanity-checks in a recent paper, and allows trustable independent reproduction à la OpenCLIP. 👍
1
2
36
0
0