This model is a basic Anima V1 minus Anima P3 LoRA extract provided as a helper to "upgrade" anima P3 merged/finetuned models:
# python extract_lora_U.py anima-preview3-base.safetensors anima-base-v1.0.safetensors P3-to-V1.safetensors
Loading anima-preview3-base.safetensors...
Loading anima-base-v1.0.safetensors...
Extracting...
100%|████████████████| 685/685 [03:26<00:00, 3.32it/s]
Stored 426/685 keys to LoRA.
Saving P3-to-V1.safetensors ...
Done!Provided as-is, you can use and merge it as you wish, no credits required.
It can be used at positive strength on a P3 finetune to catch up V1 or at negative strength when merging a style LoRA made for P3 with V1 to limit the impact.
Legal
The CircleStone Model is licensed by CircleStone Labs LLC under the CircleStone Non-Commercial License. Copyright CircleStone Labs LLC. IN NO EVENT SHALL CIRCLESTONE LABS LLC BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH USE OF THIS MODEL.
Description
Use of this LoRA must abide ANIMA license.
FAQ
Comments (6)
The extract lora thing seems like an amazing tool, how is it used? would there be a simple GUI version at some point?
It would really help save some diskspace.
The python script is "easy" to use, not sure if i'll make it a standalone GUI or a Forge extension in the future
@n_Arno I dislike using the command line somehow I never get the results I want despite writing as seen in the example, lmao. I got frustrated enough to use googles antigravity to make a simple gui, that worked well enough.
@kitekholin Since the CLI need exactly 3 parameters: original and finetuned model name and output lora name, this feel overkill 😂
But as long as it works for you, that's great! You can always consider contributing or opensource it if you'd like 😉
@n_Arno I tried it again on cmd just to see if I could, lmao. I got an error lol.
Here it is: File "E:\Git\Anima_extract\extract_lora.py", line 9, in <module>
from safetensors.torch import load_file, save_file
ModuleNotFoundError: No module named 'safetensors'
@kitekholin Ah, yes, indeed this need some python package installed (safetensors, torch and tqdm), my bad, i did not provide a requirements.txt file 😅
I'll make a GUI for sure, bundled with other tools (merge LoRAs, merge LoRA to model, extract LoRA from model)