ANY ONE WHO DOWNLOADED EPOCH 10 I MEANT TO POST 12 SO YOU SOULD REDOWNLOADNSFW - link to fantia/pixiv - <a target="_blank" rel="ugc" href="https://www.pixiv.net/en/users/17692852">https://www.pixiv.net/en/users/17692852</a><a target="_blank" rel="ugc" href="https://fantia.jp/fanclubs/299057">https://fantia.jp/fanclubs/299057</a>Would appreciate images, feedback, requests, reviews, and of course comments in the comments. Please check out my other models if you like this one.A Lora trained on the artstyle of Midori. I think it is slightly off, but sometimes it looks very close. there are two Epochs as I couldn't decide which to go with.weights 0.6 - 1 (0.9 - 1 usually looks bad, but sometimes they can look good)no trigger word. I recommend putting a phrase similar to "drawn in the artstyle of 'Lora' " somewhere in the prompt.A lot of the dataset was hypnosis/mind control, so related tags should work well, but recommend using hypnosis Loras if going for that effect.may mess with hands. is probably censored since as far as I know no uncensored art existed, but depending on your model putting censored tags in the negative might help.seems to like to make the background weird or add light like things in the background, prompting may get rid of this. might be able to get cross-section/ x-ray with this since some existed in the dataset but using a specific model or Lora is probably better same with heart-shaped_pupils.nsfw works quite well.plenty of paizuri in the dataset so seems to do decently without Paizuri Loras.optional tags: hypnosis, mind control, expressionless, ?, drool, shaded_face- END OF IMPORTANT INFO -non important rabbling belowsome of the images for each Epoch might not match as I am tired when uploading and the decision to upload 2 Epochs was a last minute decision. This was originally just going to be Epoch 9 but then I decided to mess with 10 some more and liked it. I had a whole pros and cons list for each but then scrapped it once I realized some info was wrong and most people are probably just going pick the one they like better. I was thinking of making some free poll thing so people could add and vote Loras, but most of them require money and I don't have a job so Can't do that, unless someone knows a free site. It's like 2 in the morning, should probably finish this up and set up training something elsenoticed that I accidently used the 12th Epoch in almost all of them of them. oops, will fix probably tomorrow since i'm tiredplanning to do:reverse paizuri (lying down and the "squat-like?" position)paizuri invitationvault girl/vault jumpsuit

Style - Midori

A Lora trained on the artstyle of Midori. I think it is slightly off, but sometimes it looks very close. there are two Epochs as I couldn't decide which to go with. Epoch 9 is a bit less accurate and is less unstable and follows the prompt better and Epoch 10 is more accurate but more unstable and seems to hypnosis prompts better.very NSFW oriented but can still do SFW.weights 0.6 - 1 (0.9 - 1 usually looks bad, but sometimes they can look good)no trigger word. I recommend putting a phrase similar to "drawn in the artstyle of 'Lora' " somewhere in the prompt.A lot of the dataset was hypnosis/mind control, so related tags should work well, but recommend using hypnosis Loras if going for that effect.may mess with hands. is probably censored since as far as I know no uncensored art existed, but depending on your model putting censored tags in the negative might help.seems to like to make the background weird or add light like things in the background, prompting may get rid of this. might be able to get cross-section/ x-ray with this since some existed in the dataset but using a specific model or Lora is probably better same with heart-shaped_pupils.nsfw works quite well.plenty of paizuri in the dataset so seems to do decently without Paizuri Loras.optional tags: hypnosis, mind control, expressionless, ?, drool, shaded_facesome of the images for each Epoch might not match as I am tired when uploading and the decision to upload 2 Epochs was a last minute decision

Files with SHA256: 31f265a4b469a8ff57f57875e777e1e2be86e0e3999a5c9368324ca197935d48