SD 512 cost $600k to train, SD 768 probably $1.3M and 4K would cost $10M I think to train also where do you source billions of images
AI
-
Training costs for stable diffusion models scale exponentially
By
–
SD 512 cost $600k to train, SD 768 probably $1.3M and 4K would cost $10M I think to train also where do you source billions of images
-
SD2 Trained on 768 Resolution Still Insufficient
By
–
SD2 is already trained on 768! But still too low res
-
Stable Diffusion Resolution Limitations and Future Improvements
By
–
People don't realize that Stable Diffusion is trained on tiny pics of 512x512px! Which is why you're getting so many artefacts, it's like an AI with really bad vision that needs glasses Reso will exponentially increase: the smallest details can be properly simulated then
-
SD2 768×768 Resolution Still Insufficient for Hand Detail
By
–
Training resolution is 512×512 pics, with SD2 it's 768×768 so that's too small to see hands properly
-
Startup Wisdom by Seneca: Ancient Philosophy for Modern Entrepreneurs
By
–
Startup wisdom by Seneca. Thanks for the book reco @oanaolt
-
ChatGPT Progress Beyond AGI Focus
By
–
ChatGPT is not AGI, but it’s a mistake to fixate on AGI so much that we don’t see the progress it represents.