<h2 id="hello-welcome!-e1boxf01y">Hello welcome!</h2>Here I leave a small introduction of the LoRA:They asked me for an update and since I am very fond of this mix I did the best I could, I hope you like it~ :3PS: Each Version is trained with slightly different data so choose the one you like best<hr /><h3 id="v1.0-(pony):-vwgzspk01"><a target="_blank" rel="ugc" href="https://civarchive.com/models/144033?modelVersionId=413047">V1.0 (PONY)</a>:</h3>For this update I used all the data from the first version (several illustrations that I had made raw).I also added several lip styles that I discovered out there. (well, actually the basis of this is the detailed lips that were almost never used but it's better to have more resources than none...)I added new generations and 2 versions of sketches, one is based on this style and the other on a style that I like (I don't know if you know the type of drawings that look like thin and shaky lines, well I had several compilations but I never got motivated to do it and I did it (I put it here).I also put in a mini pruned version where 100 style variations are labeled (they're all meh, because they were for a test, the idea of ​​that was to have several minimal changes to the base style without affecting it and it worked).I'm also adding a little freebie, it's the base design of my photo (nothing special).<hr /><h3 id="extras:-wh2rjxx8x">Extras:</h3>If you read this days ago, you will know that there was a mini explanation of the ease that the lora had to learn certain poses, leaving all that aside, I removed the less important triggers and left only the basic and most important one, if you like to experiment with the variations just check these out here:<ul><li>CuteMaster_PS, Sketch</li><li>CuteMaster_PS, CMPSR</li><li>CuteMaster_PS, AGGA_LG</li></ul><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/a6834622-a002-416e-bc3d-5a46ecbc2cdb/width=525/a6834622-a002-416e-bc3d-5a46ecbc2cdb.jpeg" /><hr /><h3 id="v1.0-sdxl(animagine):-cyccxp8fm"><a target="_blank" rel="ugc" href="https://civarchive.com/models/144033?modelVersionId=458106">V1.0 SDXL(Animagine)</a>:</h3>I came back again with this.This time I trained it with Animagine so it is easier to use with other sdxl models, although there are quite a few pony models now and I am quite hooked on the cartoon version.For this version I corrected some things and improved the art a little in the sketch generator (I had already done this in the SD version but it passed without pain or glory [let's cry ~]).Well... I've tried it with some checkpoints but choose the one you like best, for now I'll try to get a good lora for the mustache meme, see you later!PS: I'll see if I can finish the Lusamine lora by today (I had a problem with the labeling and I started training an unfinished version by mistake but I already solved it so now we just have to wait which version turns out better ~).<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/8729a08e-ba18-4142-9cd7-9f27a1c27f90/width=525/8729a08e-ba18-4142-9cd7-9f27a1c27f90.jpeg" /><hr /><h3 id="cutemaster_ps-or-checkpoint_type:-dyba7rw7b"><a target="_blank" rel="ugc" href="https://civarchive.com/models/457339/cutemasterps-or-checkpointtype">CuteMaster_PS | Checkpoint_Type</a>:</h3>This is a small experiment I did after learning how to generate this type of checkpoints.<ul><li>You can try it <a target="_blank" rel="ugc" href="https://tensor.art/models/728152546427828098">here</a>!</li><li>I was also thinking about using it to train characters because I like the style and I was too lazy to generate many images for it.</li><li>Actually just mix Hassaku with SD1.5<a target="_blank" rel="ugc" href="https://civarchive.com/models/144033?modelVersionId=431869">v2.0</a>.</li></ul><hr /><h3 id="v2.0-(sd):-nnl5qn6ze"><a target="_blank" rel="ugc" href="https://civarchive.com/models/144033?modelVersionId=431869">V2.0 (SD)</a>:</h3>I had planned to use the same resources that I used with pony, but after so many generations of yours and mine, I can say that I polished quite a few problems and discarded unnecessary data. This version is trained with a greater flow of faces but I greatly limited the detailed lips and discarded the "CMPSR" plugin since it is honestly not very useful and I only added it to perform some tests that I will implement in another project I am doing.So... this should be the final version for SD until further notice, regards ~<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/63814523-5075-46a5-9efa-436bea593aab/width=525/63814523-5075-46a5-9efa-436bea593aab.jpeg" /><hr /><h3 id="v1.0-(sd):-gksshhwfk"><a target="_blank" rel="ugc" href="https://civarchive.com/models/144033?modelVersionId=159970">V1.0 (SD)</a>:</h3>This style was made after changing many images with a pastel filter.The idea was to have something cute but kinky, this does not mean that the results will be hentai, but it supports that since it has enough information to not need a LoRA to know how to make a nipple. Well, enjoy! xD<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/6f81e190-5823-42c9-b55e-b21e05db7036/width=525/6f81e190-5823-42c9-b55e-b21e05db7036.jpeg" />

SD, SDXL & PONY | Cute Master_ Pastel Style

Since I didn't find PDV6 in Tensor, the seeds were not saved. (if you know my styles you will know that many examples do not have data, it is a problem with the generator I use).Although I managed to rescue the instructions and the results are great, I must emphasize that the sketch with and without color are two different styles (depending on where you add the word sketch).The "sketch" separated from the main trigger is the basic one taken from it and the other is the type that I like: (the CuteMaster_PS, Sketch that I like).This way, if you manage to generate more than four images in a row with the same instructions you can generate gif XDAnother thing, the new version gives good results on other SDXL models, but I recommend PDV6, it doesn't matter if the vae or the skip clip is missing, it's very good ~ 7w7)/

Files with SHA256: 449023fe7e08866bebddb195a119ede314c22a6bafdde5f9f504d90867021b93