
Boris Dayma — The Story Behind DALL·E mini, the Viral Phenomenon
Gradient Dissent: Conversations on AI
00:00
How Much Data Did You Tran?
The model was created by Pram o alinow to detect if an image has a warremark or not. When i trained the larger model, it reache like maybe a critical stage where suddenly it became good enough for virality. So its interesting, it already reached the threshold that interests people. Ande, there's easy ways to make it still better. The sma asked how much de did you tran? A run may be 400 million. Ah, 405 hundred milon but you know, the data is actually, is actually very important.
Transcript
Play full episode