https://github.com/diStyApps/Safe-and-Stable-Ckpt2Safetensors-Conversion-Tool-GUI

Oh ya, about...I already have it. But eh....I don't know how to run it. I am not a python guy you know... and it gives me this "ModuleNotFoundError: No module named 'PySimpleGUI", so you see.... :-(

Ok. Done. Figured it all out. In fact, there is a windows version of converter. I downloaded the python script wrongly. But also a good thing because I got to learn pip, gui and torch in 15 minutes. LOL...

Found this on another comment: "It was an individual model that I combined for about two months. Therefore, I do not remember most of the models that were combined. The last thing I remember is: HassanBlend 1.5.1.2, ProtoGen 3.4, Analog Diffusion."

So it sounds like a megamix.

I suppose not

@SG_161222 ahh, so I can post the photos generated by me on social media, without any problem, no problems with rights ?

@SG_161222 I understand, thank you

How can i train the model with my photos on colab?

can i train my own images with this model.

@SG_161222 I tried yet still it shows model.ckpt and model.safetensors [8194f84cdc] in menu. Should it show the name of the model or model.safetensors [8194f84cdc] is actually realistic vision?

@SG_161222 Oh okay thank you :))

I've found that training the model with your Face skews the model a lot, unless you only run it for few iterations. What I find more useful is to just use this model for the face generation and use the batch face replace extension to replace the poses with this model and then replace the faces with my face generated by a trained version of this model + my face.

@SG_161222 that's a great news! sure I'm waiting patiently for your next masterpiece ;)

@SG_161222 thank you, that was very helpful

@ThomasLynch checkout runpod tutorials on youtube, chanel SE software engeneering courses is a good one to checkout

@SG_161222 I took a look at those models you mentioned, they are great models, but your mix surpassed the realism of them in my opinion. Thank you.

@SG_161222 Thanks for the merge, but good to know what you mixed. Should probably add that in the description. Since some of those models cannot be used commercially.

@SG_161222 I really don't understand them much myself, but from my vague understanding, is that most models are fine because they are under the openrail or whatever its called, which allows for basically anything. However a few models like seekart and dreamlike have restrictions on what can be done with the models, mostly that you cannot sell the model. I do not remember if they restrict selling of images or not. However various models use those in merges, and then those models get merged into others. I have had that issue in my own, where I found out afterwards that some of the models used those that supposedly limit what can be done. As for your merge:

Protogen has no selling model or images, and no generation service:
https://civitai.com/models/3666/protogen-x34-photorealism-official-release

HassanBlend has restrictions of no selling model and no generation service:
https://civitai.com/models/1173/hassanblend-15-and-previous-versions

RealEdenApocalypse has the hassan merged, but the others I think are ok in it.

Those seem to require the same license, but the others seem fine, I think. I am honestly still trying to wrap my head around what is allowed with the licensing and if merging changes that. For now though I just assume that if something was merged/trained on something then it follows that license. I could be wrong but that is just my understanding.

For any of the models that fall under that grey area or cannot use area, I just use them for personal use, and just see what I can make with them. As for things I would use in projects I try to stick to ones that I can at least sell images. Hope that helps. Cheers.