Hi everyone, as always, all comments and criticisms are welcome, as they help me to continue learning!
If you like my content please don't forget to rate it! ★ ★ ★ ★ ★
➤ Info: 150-step TI trained on a dataset of 30 images
This time use a different configuration.
Vectors per token: 6
Gradient accumulation steps: 15
The rest of the training is as usual.
Here is the article where I explain step by step how to create a Textual Inversion my way:
https://civarchive.com/articles/692
➤ Based on lucillaj (photo model)
Description
FAQ
Comments (1)
Nice! I personally use 10 vectors as standard for subjects. Give it a try if you want, I think you'd like the results.
If you have to make styles, then go higher than 10 (we're talking about datasets of 50 to 100 or more pics).
That's what I've deducted from my experimentation.
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.