Excellent

#2
by moodyman100 - opened

Impressive so far. Thanks

thank you. I like version v1.2 a lot. Im working on a fl2va version of it now.

what best prompts to use?

It is the best one I found

im working on fine tuning the entire model and building a better lora version

I'm so excited to try out these loras, but I'm having some issues. Is there a place we can obtain a little more info on implementation? I'm using flux2-klein 9b-uncensored-q8.gguf clip and maybe that's not the right setup?

Because I'm using gguf, I have to split my weights, but honestly, I would rewire my whole flux workspace to get this working. Any suggestions?

I dont think thats the right clip for minimax h3

Thanks for the reply and all the work you put into making these. I did some troubleshooting and noticed a fluffy dog appeared in one of my pics, so figured that was an issue with text encoding. Sure enough, I had forgotten to wire my positive clip so it routed through the lora. After that I switched to the Power Lora Loader with some pretty decent results.

Yes, you're right about h3, sorry about the wrong thread. It was a late night.

Could you give me some advice on how to write the prompts? What kind of prompts should I use to get the best results with your LoRA?

Sign up or log in to comment