license: creativeml-openrail-m
Hello, sorry for my lousy english.
DucHaitenAIart-beta, actually 4 AI models v6.0 v8.0 v9.0 and v10.0 Version 6.0 is currently the most stable version, although versions 8-9-10 may give better results in some prompts and styles.
Each model is slightly tweaked by me to figure out which one is the best, but overall they're all a little anime oriented, a little 3D and a little realistic, and of course... have both SFW and NSFW.
but since it's all beta, i need to try it many times to get research data, and it's too difficult for me alone with a graphics card with only 4gb vram, so i need help from all guys, try it out and post it on all social networks with the tag duchaitenaiartbeta (the version number you use). Or send me directly if you find any problem with the AI model
Try it as much as you can with every prompt and aspect ratio
A little reminder: Sometimes a negative prompt is more important than a prompt, Please write a long and detailed negative prompt, but don't be too lengthy and redundant, preferably concise and just enough to express all you want to express.. The setting i usually use is 28 steps, the sampling method is Euler a, CFG Scale 11, for me it works best, but maybe it's still not the best setting, so I need people try all different settings
please support me by becoming a patron, please i need a more powerful computer 😭😭😭 http://patreon.com/user?u=83557525
Some test:
