ArtiVisionXL (Experimental)版本ArtiVisionXL_mixV1 (ID: 601627)

ArtiVisionXL (Experimental)版本ArtiVisionXL_mixV1 (ID: 601627)

Here comes my first Checkpoint ...

this was more kind of an experiment, since I just wanted to create a checkpoint for my own usage. I have no experience in merging models or with comfyUi (which I used for the merge). So I had to force myself to use this "interface", to which I still have a very minimal to a non-existent affection. ? ... But I´m really surprised of how well it works. So I decided to upload it on Civitai.

This is a merge of the following Models:

Juggernaut XL V9 + RDPhoto2-Lightning_4S (everybody knows the quality of Juggernaut models)

Pixel Alchemy Hyper V2.0 (an extremely underrated checkpoint and a personal favorite of mine, so please, go ahead and also try that one out! ...)

xl_more_art-full /xl_real /Enhancer Lora (which also is very well known and extremely estimated)

Here are some tips and suggestions!

Yes, I merged a lightning model with a hyper model, and I was not sure if that would work ... but it does ... if you follow the suggested settings ?

I M P O R T A N T !!!

Suggested Settings:

Sampler:

  • Euler a (from testing, and to my surprise this one works best!)

  • Euler (also really good results)

  • Heun, DPM2, DDIM (seem to also work, but the results are clearly lacking in quality)

  • all other Samplers within A1111 don´t work properly or the result is poor.

  • I have not tested any of the additional Samplers, which are available with comfiUI, ... I don´t use comfiUI for image generation, but feel free to do some testing of your own

the inflexibility in sampler usage might be the only disadvantage of this checkpoint

Steps: 12-25 Steps for best ratio between performance and quality

CFG: 2,5 - 4,0

VAE: baked in

Hires. fix values I use:

  • 25 Steps

  • 0,2-0,4

  • 4x-UltraSharp-upscaler (my choice) or any of your choice

  • upscale by 1,5

Additional Information and Usage Tips:

As you can see by my sample images, this checkpoint is capable to depict pretty everything you can imagine. However, I´m by now also pretty comfortable with prompting. So, if you don´t get the same results first try, don´t be discouraged and don´t just blatantly copy. Get inspired, make some changes and adjustments, ... put in a little love and effort. It´s much more rewarding, trust me! ?

Please, don´t upload rough XXX stuff, it won´t show up here, I´m sorry but this checkpoint isn´t meant for that type of "art".

All sample images were created with A1111, without the use of any loras, no inpainting, no further upscale method (beside the use of hires fix), and some of them even without the use of Adetailer (which can often enhance but also destroy the overall harmony).

If you have further questions, feel free to ask. I´ll try to answer asap. Please keep in mind though, that I have a full time job and other obligations (like most of us have) ...

Now, have a lot of fun with my checkpoint and consider to leave a review/like. ???

描述:

训练词语:

名称: artivisionxl_Mixv1.safetensors

大小 (KB): 6973384

类型: Model

Pickle 扫描结果: Success

Pickle 扫描信息: No Pickle imports

病毒扫描结果: Success

ArtiVisionXL (Experimental)

ArtiVisionXL (Experimental)

ArtiVisionXL (Experimental)

ArtiVisionXL (Experimental)

ArtiVisionXL (Experimental)

ArtiVisionXL (Experimental)

ArtiVisionXL (Experimental)

ArtiVisionXL (Experimental)

ArtiVisionXL (Experimental)

ArtiVisionXL (Experimental)

ArtiVisionXL (Experimental)

ArtiVisionXL (Experimental)

ArtiVisionXL (Experimental)

ArtiVisionXL (Experimental)

ArtiVisionXL (Experimental)

ArtiVisionXL (Experimental)

ArtiVisionXL (Experimental)

ArtiVisionXL (Experimental)

ArtiVisionXL (Experimental)

ArtiVisionXL (Experimental)

资源下载
下载价格VIP专享
仅限VIP下载升级VIP
犹豫不决让我们错失一次又一次机会!!!
原文链接:https://1111down.com/1059366.html,转载请注明出处
由于网站升级,部分用户密码全部设置为111111,登入后自己修改, 并且VIP等级提升一级(包月提升至包季,包季提升到包年 包年提升至永久)
没有账号?注册  忘记密码?

社交账号快速登录