Cant we further tune this model wit dreambooth, or hypernetwork or embeddings?
I want to teach my face on this model is that possible?
Yes, I did it and it is great! Did it on my first attempt of using Analog Diffusion in fact, without trying the original first. Which perhaps I should do to compare the quality of the results.
Excellent. Which version of Dreambooth did you use? The Last Ben has been returning errors for the last few days…
Hmm, I'm not sure where I can see the version number?
It's not a version number. It's the name of its developer. There's The Last Ben, Joe Penna, and ShivamShrirao versions under Google Colabs. It's visible in the web address as well. There's the Automatic1111 repo extension. These are the most popular options.
It's not a version number. It's the name of its developer. There's The Last Ben, Joe Penna, and ShivamShrirao versions under Google Colabs. It's visible in the web address as well. There's the Automatic1111 repo extension. These are the most popular options.
Hey mate, would you mind to share your training code?
I've trained my face with ShivamShrirao's Dreambooth on the second time with this model, but none of them couldn't keep the inheritance of Analog Diffusion model.
It's not a version number. It's the name of its developer. There's The Last Ben, Joe Penna, and ShivamShrirao versions under Google Colabs. It's visible in the web address as well. There's the Automatic1111 repo extension. These are the most popular options.
Ahh, sorry. Thought I mentioned that but it must have been in another discussion. I use TheLastBen. Dreambooth for training and Automatic1111 now lately for prompts.
Could someone please share the basic steps (for training)? What did you use? How many pictures? What settings etc?
Thanks in advance.
I've been trying on TheLastBen's repository, but I'm not getting any decent results, even with hq images and different "optimal" settings.
I've been trying on TheLastBen's repository, but I'm not getting any decent results, even with hq images and different "optimal" settings.
ye it ,is hard to get good results due to pruned weights i think
but you may try my tutorials
1: https://www.youtube.com/watch?v=mnCY8uM7E50
2: https://www.youtube.com/watch?v=2yGGorOxtbA
playlist of tutorials for those who wonders : https://www.youtube.com/watch?v=mnCY8uM7E50&list=PL_pbwdIyffsmclLl0O144nQRnezKlNdx3