1 Answer
- Newest
- Most votes
- Most comments
2
Hey there, looks like you are trying to train your stable diffusion model on a multi class training set. Unfortunately this is not a feature in SageMaker JumpStart just yet. There is a way to get around this however which follows the strategy of
- Fine-tune the model for Subject A.
- Fine-tune the resulting model from Step 1 for Subject B.
- Generate images of Subject A and Subject B using the output model from Step 2.
If you’re after more information / an example, check out this blog and scroll down to fine tuning considerations. It goes through the process of training a stable diffusion model with one class then runs through how it has trained it on multiple classes and the associated strengths and weakness of both approaches. Hope this helps!
answered a year ago
Relevant content
- asked a year ago
- AWS OFFICIALUpdated 6 months ago
- AWS OFFICIALUpdated a month ago
- AWS OFFICIALUpdated a year ago
- AWS OFFICIALUpdated 17 days ago
Thanks for this