Trained on Aurora XL base[pony XL] v1.0 https://civarchive.com/models/301957/aurora-xl-basepony-xl. It is a pony based checkpoint that is great for style LoRA training.
Regarding LoRA file size:
A higher DIM results in a bigger size.
Training with a lower DIM (for example 32) takes slightly less amount of time but gets worse in learning artist style. If you need to reduce LoRA size, try "remake dimension" using extension SuperMerge.
Here is comparison of:
1. anmiXL_V1 (870MB)
2. anmiXL_V1 remake dimension to DIM32 (217MB)
3. Same dataset trained with DIM32 natively (217MB)
https://civarchive.com/posts/1379092
I personally feel "remake dimension" is better than training with DIM32 natively.
Description
FAQ
Comments (4)
Does XL need a dedicated lora for each checkpoint? Oh, my poor hard drive -- translation software
Animagine V3 and ponyDiffusion are too different. LoRAs trained on Animagine V3 won't work on ponyDiffusion. I'm now moving to ponyDiffusion and I like Aurora XL most.
dude, lora 800mb is too much. even 200mb is big but 800 is just too much. why is this so big?
A higher DIM results in a bigger size.
Training with a lower DIM (for example 32) takes slightly less amount of time but gets worse in learning artist style. If you need to reduce LoRA size, try "remake dimension" using extension SuperMerge.
Here is comparison of:
1. anmiXL_V1 (870MB)
2. anmiXL_V1 remake dimension to DIM32 (217MB)
3. Same dataset trained with DIM32 natively (217MB)
https://civitai.com/posts/1379092
I personally feel "remake dimension" is better than training with DIM32 natively.
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.