Shinoa

Hello everyone!Well this just contains some of my models and projects that I will be working on soon.ListShinokira - v1.0 - CheckpointKibouMirai - v1.0 - CheckpointYumehana - v2 - CheckpointShineStyle - v0.5 - LoraLuxHyphen - v0.1 - LoraFukkatsuYoake - v1 - LoraYuukyuuNozomi - v3 - LoraSaisei - v1 - LoraGurenSutorīto - v3 - LoraReimei - v1 - LoraCheapPro - AIToolsCHRISTMAS❄️- AIToolsButton❄️- AIToolsSimple - AIToolsHistoryCheckpoint1. ShinoKiraFeb 9, 2025 : Upload Shinokira - v0.1, My first checkpoint, the details might be pretty good or it might look overdone, but the double-highlight issue on the hair annoys me a bit.April 21, Reupload to change the name of the Model file.May 6, 2025 : Upload Shinokira - v1.0, May be enough, but since the base used is Illustrious XL - V2.0/Stable which is trained to use narutal language prompts, from some of my (temporary) testing, if using natural language prompts the results are sometimes better when compared to description prompts which feel more rigid (depending on the prompt entered).Jan 8, 2026 : Upload Shinokira - v2.A and Shinokira - v2.B. I noticed that many people liked this model, so I decided to update it with the latest dataset. I used Illustrious XL - V1.0 as the base model, but there are also differences in the dataset due to several considerations. The difference between v2.A and v2.B lies in the dataset used. For v2.A, it is the same as Shinokira - v0.1 but with a different base and the same dataset. Meanwhile, for v2.B, the base is the same as v2.A but without the added dataset.2. KibouMiraiMay 6, 2025 : Upload KibouMirai - v1.0, My second checkpont, it may be said that ShinoKira and KibouMirai look the same but there is a slight difference between the two, which is from the training process and ShinoKira has a unique style while KibouMirai only has a basic style.Soon : Maybe I will try to use Illustrious XL base - V1.0.3. YumehanaOct 4, 2025 : Upload Yumehana - v1, My third checkpoint, using the Raehoshi illust XL - v6 model as the basis for my model and adding a new style dataset as its base style, but the results were not as I expected, so I will try to retrain it until it matches my dataset.Nov 22, 2025: Upload Yumehana - v2, I updated the base using ‘Raehoshi illust XL - v7.1’, but I think the results are still the same even though I added my own dataset to change the style. I will keep trying to update it.Lora1. ShineStyleMar 4, 2025 : Upload ShineStyle - v0.1, Issue has no effect when used and there is a leak.Mar 25, 2025 : Retrain ShineStyle.Mar 26, 2025 : Upload ShineStyle - v0.5, Issue is still not too visible the difference from the previous version and a little leak.2. LuxHyphenApr 7, 2025 : Upload LuxHyphen - v0.1, it works but needs a little improvement.April 21, 2025 : Reupload to change the name of the Model file.3. FukkatsuYoakeJul 3, 2025 : Upload FukkatsuYoake - v1, This Lora style uses the same dataset as Lora ShineStyle with a few additional datasets and better tagging.4. FukkatsuYoakeNov 26, 2025 : Upload FukkatsuYoake - v1, this is the first lora from my experiment. I will try to update it until I get a good lora result.Dec 16, 2025 : Upload FukkatsuYoake - v2, this is the second result of the experiment, which differs in terms of dataset size and steps.Dec 20, 2025 : Upload FukkatsuYoake - v3, Ttis is the third result of the experiment, which differs in terms of dataset size and number of steps. In my opinion, it is quite good and close to the dataset.5. SaiseiDec 9, 2025 : Upload Sansei - v1, this is the second lora from my experiment, used Illustrious XL - V1.0 as the base model. I will try to update it until I get a good lora result.Dec 28, 2025 : Upload Sansei - v2, I updated this lora style, still using Illustrious XL - V1.0 as the base model, and I like the results.6. GurenSutorītoJan 3, 2026 : Upload GurenSutorīto - v1, my third latest lora style experiment, using the Raehoshi illust XL model as its base. I'm still confused because the results are not as expected, but I will continue to try to update it.Jan 17, 2026 : Upload GurenSutorīto - v2, I updated this lora, still using the Raehoshi illust XL model as its base. There are slight differences from v1, because the Raehoshi illust XL base is so strong that it makes it a bit difficult to use as a base for lora training for me, but it might work because it's almost identical to the dataset.Jan 25, 2026 : Upload GurenSutorīto - v3, I tried using a different base, I used the RouWei base epsilon version, because it was talked about a lot and made me interested in trying it. But maybe the result still confuses me, whether it's enough or not.7. ReimeiJan 23, 2026 : Upload Reimei - v1, Lora style to the fourth result of my experiment, I tried using RouWei as the base model epsilon version. The results show a slight difference when experimenting with the base and with Lora. I will try to update this Lora.AI Tools1. CheapPro : -2. CHRISTMAS❄️: -3. Button❄️: -4. Simple : Same as AITools CHRISTMAS❄️but this can be updated.6. I tried to create a Lora WAN and AI Tools before, but it seems that I am not skilled enough to create Lora WAN 😅Soon1. Checkpoint : I try to keep developing the existing models.2. Checkpoint 1 : Actually, I am trying to create a checkpoint from scratch with the latest dataset danbooru for April 16, 2025. However, I encountered many obstacles when I tried it, but I will keep trying.3. Checkpoint 2 : Maybe if I succeed in training model checkpoint 1, I will make another one with the addition of the anime style that I like, so just wait and see.4. Lora Cosplayer : Actually, I've already tried to create a cosplayer lora. When I tried it, the results were somewhat smooth in some parts, but there are still some things I need to consider before releasing this cosplayer lora..5. Lora : I need a specific lora, maybe I'll try to train one.6. Lora : There will be new style models coming.7. Lora : Maybe in the future I will try to create characters from anime or games.Additional notesHere I conducted several experiments to create a model. I would appreciate it if you could try it out and provide feedback in the comments section.If you have any questions, suggestions or criticisms for me, you can leave them here Leave feedback.*I don't know why this article was recommended again on the Explore channel by Team TA. Maybe they instructed me to update this article 😅😂 - Thank you, Team TA 😆See you again in the next update. Bye Bye! ☺️