Description
FAQ
Comments (5)
It might be a cheap shot, but for completion with these great cyber-civilization Loras, may I suggest an African-inspired one as well? A la Wakanda :)
🔥👍
Your LoRAs are getting better and better.
In most new ones overfitting burn is gone when set to 1.
And they do not negatively impact other embeddings when combined.
How many images in this dataset and how many repeats per image?
Thank you very much for your support and I am very happy to hear that they work well! In the latter I increased the steps because depending on whether I use Fine-tuning or not. In this case it's not fine tuned, the steps are around 240 per epoch, 10 repeats of the dataset. This model I used a very small dataset of 42 images. This is because they were created on purpose
@hk4life Thanks for the info :) Very interesting concepts you are coming up with. Cyber Egypt is amazing as well.
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.