Let me know what you think and any suggestions on who to see next.
If you like my work and wanna show me buy me a coffee
Description
FAQ
Comments (4)
Great model. Worthy of a star rating.
Congratulations.
Its about time someone KNOWS what they are doing and how to do a model for SD.
Thank you and congrats on a well done model.
Very useful for when using Checkpoints which cannot create her face,
---
If you do not know, using the Checkpoint Realistic Vision version 1.3 , you can create a super-realistic and well done face (and the same body) of Eva. It all depends how well you know how to create a proper Prompt for doing Photo-realistic AI-images.
Using this checkpoint you can create over 500+ famous people without the use of any Lora/Lycoris or T.I..
Wow thanks!
Yea I use realistic vision a lot. Its probably the best one to be "real". More so I've been using it to create some Lycoris (will be having more of those upload later this month).
@frankyfrank2k Be careful. That is NOT what I meant. I said to Use Realistic Vision version 1.3 to create more than 500+ famous persons WITHOUT using LORA or Lycoris or T.I.
...Using Realistic Vision version 2.0 will NOT make the same number of famous faces.
No other Checkpoint comes as close to creating photo-realistic pictures as version 1.3.
I know, because I have been testing it and using it since December 2022.
Here some examples for you to see and know ....
https://i.imgur.com/VuiQgOU.jpg
https://i.imgur.com/4CHOqDC.jpg
https://i.imgur.com/SuuGCpx.jpg
https://i.imgur.com/3jziP2r.jpg
https://i.imgur.com/XgdQVXX.jpg
-
ALSO Very Important = NEVER train your image face models on any checkpoints. Otherwise it may give users problems if they do not use the same checkpoint.
The BEST way to train a LORA/Lycoris/Textual Inversion is to use the original Stable-Diffusion checkpoint which is 1.5-pruned.ckpt or the 1.5-pruned.safetensors .
https://huggingface.co/runwayml/stable-diffusion-v1-5/tree/main
That way the trained model can be used in ALL other checkpoints which are created for Stable Diffusion,
without problems. Otherwise, the user will have to also download the same checkpoint which the Lora was trained on, so the face can look better.
Do you understand the process ?
@maxhitman Ah ok I get ya.
Yea all TI I have done are done with base SD 1.5. Like you said it allow it to be flexible with other checkpoints as long as their base is SD 1.5.
Though I am curious of running a couple on RealisticVision just to see how those turn out. For giggles.
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.