Checkpoint merge test
Description
FAQ
Comments (15)
Thanks, this is the 3rd model that I found that works well with my loras! Would you make fp8 versions too?
sorry i dont know how to do it. If you can guide me I will be pleased to do so.
@netoschka I don't have the specs to finetune models locally, but supposedly you can use ComfyUI's ModelQuantizer to "resize" the model. a FP8 model would be around 6G in size so it runs on lower VRAM too.
Not sure where you finetuned your model though, just found this wiki link on how to start: https://deepwiki.com/lum3on/ComfyUI-ModelQuantizer/1.1-getting-started
I can see the preview image on the model list. But when I open the page there are no images. Maybe its a Civit bug?
I don't know what's going on with Civitai, they've been acting badly lately.
where is V9.0?
Could you post in the normal 6.7GB size pls, 13GB is too much to DL and no space on my hard drive. Thanks (also takes like half hour to download a 6.7GB model from CivitAI).
I would like to know how to do that. do you have any python code to do it?
@netoschka If you use Comfyui it is really easy to do with a connection from the model (from the checkpoint loader) to the built-in ModelSave node, which will save the unet, that is the diffusion model without the text encoder and vae, if you run it. I think it saves it in the diffusion_models folder.
DL Broken
not downloadable. download alway crashes at 60-80%.
downloadmanager not usable, resume download not possible.
no idea why this happens.. might be a civitai problem?
Why is version 2 newer than version 5? Why didn't you just make it version 6?