SD3.5 - Emma Watson
In my opinion, SD3.5 It is not yet as good as Flux for realism but Finetunes should make it way better over time...
Training settings and especially generation settings are NOT locked in yet - Results may vary greatly, I will be updating with new versions as I experiment further...
Thank you for understanding and let me know if you know good generation settings for SD3.5
Want a Custom/private LoRA? Get it here : Ko-Fi Commission
Prompts in showcase images
Enjoy!
Leave me a review so it can improve!
Description
FAQ
Comments (8)
Results look pretty good ! How many epochs ? How long did it take ?
How are you training SD3.5 please? Which tool(s)?
Hello,SD 3.5 requires more VRAM than Flux to Train?
What tool are u using?
ss_base_model_versionsd_1.5
??????
Does the base model recognize celebs at all?
Great work making it out so quickly. I have great hopes for SD3.5 fientunes!
Thank you for uploading this LoRA since I switched over to 3.5 this week away from FLUX, I wanted to start testing some LoRAs on it before I train my specific LoRA for my own generations. I just tested it and it looks great as Emma Watson is truly generating on it, so all is working well for me on my end.
Are you planning on uploading more character LoRAs to this site?
anybody got a workflow for sd3.5 large with lora? cant find any online and i think this may be the only character lora on civitai for this model lmao. just want to test it.