novelchef

NovelAI Tutorial + Behind the Scenes of OP 15 + CC 3 Schedule Update (Volume warning. My mic and the AI spike sometimes so keep the volume low.) It's here boys and girls, and it's a big one- about 85 minutes long. I go over almost every aspect of my process through NovelAI, Pixiv, and a bit of Photoshop while showing you most of the 3000 images that didn't make the cut for OP 15. Honestly, as a tutorial for NovelAI it's kind of mid, and I acknowledge that I probably glossed over some things I should've taken more time on. I could have talked more about tag engineering, different models and samplers, editing steps/noise/strength of images etc., but I knew the video was going to drag so I skipped a lot. Even so, I end up kind of dragging it out a bit especially while I'm showing OP 15 images, so I recommend setting the video speed to 1.5 speed or skipping to the parts you want to see. Also, I included a ZIP file with all of the Cindy images that didn't make the cut (it's 1700 images- youve been warned.) The video is 18+ and I use an AI voice of an anime character to kind my identity. (10 points to whoever guesses who it is.) This is my first time making this kind of thing, and while I censored plenty of parts there is a good chance some sensitive data is there, please don't dox me lol. Timestamps are below. 2:40 NovelAI Tutorial 11:00 Enhance/Variations/Seeds 20.43 Inpainting- The Better Option 24:00 Use Ai to Change any Existing Image 30:00 Pixlr Stuff 35:28 Photoshop Text Boxes 37:02 OP Part 15 44.40 Cindy scenes 56:55 Vibe Transfer 106:23 Explicit scenes CC is on the way, more likely to come out Thursday since I have like 4 doctor's appointments this week and the video ended up taking me a bit longer than I thought with getting the various software setup. PS: Make sure to use the tag "NSFW" when trying to generate explicit images! The link is in the locked text below and contains both the How-To video and a ZIP of all the unused Cindy images.

Published: October 22nd 2024, 5:34:19 am

PreviousNext

NovelAI Tutorial + Behind the Scenes of OP 15 + CC 3 Schedule Update (Volume warning. My mic and the AI spike sometimes so keep the volume low.) It's here boys and girls, and it's a big one- about 85 minutes long. I go over almost every aspect of my process through NovelAI, Pixiv, and a bit of Photoshop while showing you most of the 3000 images that didn't make the cut for OP 15. Honestly, as a tutorial for NovelAI it's kind of mid, and I acknowledge that I probably glossed over some things I should've taken more time on. I could have talked more about tag engineering, different models and samplers, editing steps/noise/strength of images etc., but I knew the video was going to drag so I skipped a lot. Even so, I end up kind of dragging it out a bit especially while I'm showing OP 15 images, so I recommend setting the video speed to 1.5 speed or skipping to the parts you want to see. Also, I included a ZIP file with all of the Cindy images that didn't make the cut (it's 1700 images- youve been warned.) The video is 18+ and I use an AI voice of an anime character to kind my identity. (10 points to whoever guesses who it is.) This is my first time making this kind of thing, and while I censored plenty of parts there is a good chance some sensitive data is there, please don't dox me lol. Timestamps are below. 2:40 NovelAI Tutorial 11:00 Enhance/Variations/Seeds 20.43 Inpainting- The Better Option 24:00 Use Ai to Change any Existing Image 30:00 Pixlr Stuff 35:28 Photoshop Text Boxes 37:02 OP Part 15 44.40 Cindy scenes 56:55 Vibe Transfer 106:23 Explicit scenes CC is on the way, more likely to come out Thursday since I have like 4 doctor's appointments this week and the video ended up taking me a bit longer than I thought with getting the various software setup. PS: Make sure to use the tag "NSFW" when trying to generate explicit images! The link is in the locked text below and contains both the How-To video and a ZIP of all the unused Cindy images.

You May Also Like...

PFK Finale Schedule and other updates Part #2: I'll be taking a week "off" in July to get a new PC up and running. Also, I'll be using this time to learn stable diffusion and try out different combinations of models and Luas. I'm hoping that I can get multiple models up by the end of the week that can create the kind of content NovelAI Diffusion is bad at, such as cartoon stuff or absurdly high-res images involving multiple characters. It's not gonna be a shortcut or replacement to NovelAI diffusion or anything (unless my PC runs it really fast) but I might use it more and more for intricate lewd scenes and the like. The potential for saving time is huge and I'm hoping I can learn some more tricks in the week to make up for the time spent. Once again, thank you guys sooo much for the support. I would never have been able to afford a new high-end laptop without you and my old one is quite literally falling apart. (Fan issues made it run so hot that the plastic chassis started cracking at breaking off at certain joints.) Know that the support you guys give directly goes into improving the quality of the content (and hopefully the speed too!) Also for fun, here's a side-by-side of the first image I ever mass-generated and edited compared to a new one I made a few days back. I included an image of Milly to show what the original outfit is supposed to look like. AI has come a long way since then, and so have I for sure! (Also holy smokes I used to be so proud of that earlier Lelouch edit hahaha. Like, that took forever to make look even remotely passable and the face/hands/background is so bad. As my first ever edit though I guess I'll cut it some slack.)

PFK Finale Schedule and other updates Part #2: I'll be taking a week "off" in July to get a new PC up and running. Also, I'll be using this time to learn stable diffusion and try out different combinations of models and Luas. I'm hoping that I can get multiple models up by the end of the week that can create the kind of content NovelAI Diffusion is bad at, such as cartoon stuff or absurdly high-res images involving multiple characters. It's not gonna be a shortcut or replacement to NovelAI diffusion or anything (unless my PC runs it really fast) but I might use it more and more for intricate lewd scenes and the like. The potential for saving time is huge and I'm hoping I can learn some more tricks in the week to make up for the time spent. Once again, thank you guys sooo much for the support. I would never have been able to afford a new high-end laptop without you and my old one is quite literally falling apart. (Fan issues made it run so hot that the plastic chassis started cracking at breaking off at certain joints.) Know that the support you guys give directly goes into improving the quality of the content (and hopefully the speed too!) Also for fun, here's a side-by-side of the first image I ever mass-generated and edited compared to a new one I made a few days back. I included an image of Milly to show what the original outfit is supposed to look like. AI has come a long way since then, and so have I for sure! (Also holy smokes I used to be so proud of that earlier Lelouch edit hahaha. Like, that took forever to make look even remotely passable and the face/hands/background is so bad. As my first ever edit though I guess I'll cut it some slack.)

Other Projects

Other Projects

Issue Update: So good news and bad news... I ran all the tests, and found out that the issue was with the V2 NovelAI model. It uses different servers than the V3 or something... but it's strange... Sometimes the V2 worked perfectly fine while the V3 was acting up, but it must've just been a coincidence. Anyways, so the bad news is that I think I'll have to switch to the V3 model for the office story just like how I did with Gyaru, which is really unfortunate as I liked how consistent the V2 model was in generating the main cast. (Unlike Gyaru where the character consistency was all over the damn place.) Meaning that Conor might start to look a little different halfway through the release, and the resolution will change as well to a larger size. I'm sorry if this breaks your immersion... I'll compile and release what I have at 10pm PST, messaging a link to those in the higher tiers, but just know that what I have is still mainly setup. The real feminization stuff is slated for the last 6 pages or so of this release, so I wouldn't recommend upping your sub just to see the first 15-20 pages of setup. Tomorrow I should be able to release it in full.

Issue Update: So good news and bad news... I ran all the tests, and found out that the issue was with the V2 NovelAI model. It uses different servers than the V3 or something... but it's strange... Sometimes the V2 worked perfectly fine while the V3 was acting up, but it must've just been a coincidence. Anyways, so the bad news is that I think I'll have to switch to the V3 model for the office story just like how I did with Gyaru, which is really unfortunate as I liked how consistent the V2 model was in generating the main cast. (Unlike Gyaru where the character consistency was all over the damn place.) Meaning that Conor might start to look a little different halfway through the release, and the resolution will change as well to a larger size. I'm sorry if this breaks your immersion... I'll compile and release what I have at 10pm PST, messaging a link to those in the higher tiers, but just know that what I have is still mainly setup. The real feminization stuff is slated for the last 6 pages or so of this release, so I wouldn't recommend upping your sub just to see the first 15-20 pages of setup. Tomorrow I should be able to release it in full.

Quick Update

Quick Update