This LoRA was trained on 800k images of hyper sized anime characters. It focus mainly on breasts/ass/belly/thighs/fat. This dataset is a subset of the larger hyperfusion dataset, but filtered down to body shape/size related images only.Recomendations:<ul><li>Dora/LoRA strength: 1.0 (DoRA's work in most WebUI's by now)</li><li>Resolution: ~1024px</li><li>samplers: any sampler anima supports (should be most)</li><li>in v10+ you can push the lora weight more than in v9, so do that if your concepts are not working as well as you like.</li></ul>Uploaded 1.4 million custom tags used in hyperfusion <a target="_blank" rel="ugc" href="https://huggingface.co/datasets/thojm/hyperfusion_classified_tags_export">here </a>for integrating into your own datasets (eventually ill update it one last time with my latest changes)v11 Anima_preview Release 2026/05/16:<ul><li>This LoRA was trained on the original preview version of Anima (not the newly released base_1.0), however it seems to work with base_1.0 with some minor concept lost, so you can use whichever you want.</li><li>This model is technically not done training, but I want to release what I have so far, as I switch over to training on Anima_base_1.0 (5 epochs so far)</li><li>Ive added some new concepts and improved a number of tags, as well as having about 80% of the data captioned by qwen3, so I train on both tags and captions now. (see optional data for tag guide)</li><li>Its going to be a few months to continue training on base_1.0 so ill probably go silent again as it trains. I expect it won't get really good until epoch 10+, so maybe half way there (if I don't decide to start from scratch with base_1.0, we'll see).</li><li>I train with "@artist" format, so prepend @ for artist tags in v11 just like anima does</li></ul>v10 Noob_vpred Release 2025/07/29:<ul><li>Did you guys think I disappeared? Nope, just hopelessly training a model with a frozen text encoder for 7 months.</li><li>This new DoRA has the same concepts you are used to by now, but with a few new concepts as usual. Also 200k more images than v9.</li><li>This version is trained on NoobAI_Vpred, so there is no guarantee it will work with anything else. Especially not on non-v_pred models.</li><li>Wanted to try training with the Text Encoder frozen one last time. Also decided to stick to it no matter how long it took. And now I can definitively say I will be including TE in future models just for the sake of time. It works, but its way too slow for my setup.</li><li>Use the tag list in v9 for now, until I get around to building the new one with the small number of new concepts.</li><li>This one should handle concepts a little better than v9_sdxl, and is less prone to exploding gradients as well.</li></ul>v9 Pony Release:<ul><li>This model has been training for over 2 months now, but since Flux dropped, I decided to release what I have so far to free up a GPU. Technically it should have trained for longer, but I'm impatient, and some of you are probably tired of waiting anyway.</li><li>The tags are mostly the same as the last v8 release for SD1, with a few new additions like blob content for example. See the tag.csv for more in "Training Data".</li><li>Pony is a little tricky to train on, so I was experimenting a lot with this model. Because of this you should try to keep the DoRA strength near 1.0. Anything above 1.1 tends to explode. (weight regularization like scale_weight_norms is critical for training on pony, fyi)</li><li>To keep training time reasonable I trained at 768x768 resolution initially, and had planned on finishing up training with 1024px resolution, but then Flux happened. The results still seem reasonable.</li></ul> I put plans and progress here every now and then.<h2 id="changelog-article-link"><a target="_blank" rel="ugc" href="https://civarchive.com/articles/3320">Changelog Article Link</a></h2>

hyperfusion Anima_preview LoRA 800k images

This was a test run on the original Anima_preview model, similar dataset and training params to my last release. It should still work with anima_base-1.0 but there will be some concept loss until I re-train this on anima_base-1.0The only major training differences from v10 were:<ul><li>use LoRA-Locon instad of DoRA, since lycoris didn't support anima just yet (it does now though).</li><li>Fixed caption/tag training. The last model was accidentally using captions 70% of the time instead of tags. It is now tags 70% of the time, and captions 30% of the time.</li><li>200k more images</li><li>Using AdamWSchedulerFree instead of Adamw8Bit so I didn't have to fiddle around with a scheduler.</li><li>Unet lr 8e-4</li></ul>

Files with SHA256: 78682c1438671e62c02a4ebae445e38d746245c9c3e99184175130e6ac9c86ef