We're a Day-0 launch partner for MiniMax H3, a next-gen open-weights multimodal video model - and it's already in the generator. No waitlist. Open weights land in a few days.

MiniMax H3 is live: open-weights video, generate now

Alphamon,Dorugoramon and Dexdorugoramon from Digimon as three separate models. Included are multiple checkpoints for different base models, though the quality of the 1.5 versions is questionable and they require a decent amount of prompting to produce good results. I would recommend reading my version notes for each model version since they go into detail on what the model can produce.

Dexdorugoramon trained on anima. This one turned out very well, the character is consistent with fairly minimal prompting. The trigger word is just "dexdoru", if it doesnt generate the wings properly you can add "multiple alternating red feathered wings and black bat wings". I recommend a strength under 1.0, 0.8-0.9 seems to be the sweet spot.

Dexdorugoramon Anima

Anima version trained on the same dataset as prior version. Seems to generate the character with little effort, though the fingers can be inconsistent. The trigger word is "dorugoramon", but I recommend adding "wings, horns,tail" when doing full body shots as sometimes it leaves out those. It seems to work best with a strength in the range of 0.7 to 0.9

Dorugoramon Anima

Illustrious version of the dexdorugoramon model trained to be a lower file size. It seems to be as good as the other versions, though you do need to crank up the strength up to 1.2-1.3 depending on the checkpoint your using for best results. The trigger keyword is just "dexdoru", but other trigger words help, especially with the wings.

DexDorugoramon Illus V2

Flux version of the dexdoru model trained on the same dataset, but with simple sentence captions in addition to tags. A bit inconsistent, especially at lower CFG values (I find 3.5 provides the best results), and negatives definitely help. The core trigger word is "dexdoru", though you may need to be a bit more descriptive of its overall design if you get garbled results. The model seems to work at a weight of 0.8-1, with higher weights being better if you raise the step count to around 40. I have included a prompt that seems to generate the most consistent full body image next to the trigger word if you have any issues.

Dexdoru Flux

Illustrious Version of the Death-X-Dorugoramon Model, trained locally on about 17 images for 50 epochs with sentence style captions. This version seems to work best on base illustrious derived models and requires some additional prompting on NoobAI based models, but it will still work. It seems to function in a variety of styles and models, though it struggles a bit with realism. The trigger keyword is simply "dexdoru", but additional keywords help, and I will list them in the trigger words. It seems to work well at a LORA strength around 1, but you can raise and lower it as you see fit.

DexDorugoramon Illus

Illustrious version of my Dorugoramon model trained on the same data-set as the other versions on civitai's training service. I tested it in a variety of checkpoints and styles and it seems to work consistently across all of them. It works best at a strength of around 1.0, though if you want to alter the character significantly, you can lower it to 0.8-0.9. The trigger word is just "dorugoramon", but depending on the style, keywords like "horns,yellow eyes,red forehead gem,wings,tail,dragon" may help bring out the character more.

Dorugoramon Illustrious

Flux version of my dorugoramon model trained on the Civitai training service. This version was trained on booru-esque captions but also accepts sentence prompts quite well. "Dorugoramon" is the trigger word, though it helps to describe the subject in more detail.

Dorugoramon Flux

XL version of alphamon. Seems to be unable to generate oryuken and doesnt generate the cape by default. Otherwise it seems to work with most models and styles. I recommend using a lower strength if you want to try giving him wings.

Alphamon SDXL

XL version of the model trained on BLIP style captioned images at 1024 resolution for a full 8 epochs. Seems to produce higher quality images by upscaling than the low training version, but seems to produce images with a lot of artifacts when generating at high res on turbo (of course upscaling is still fine on turbo). Seems to be flexible with a decent amount of styles as well. I find adding "wings, tail, horns, forehead jewel" to the prompt can help at lower strengths.

Dorugoramon SDXL V2

Base SDXL dexdorugoramon model. Seems to be fairly consistent at a variety of strengths while also functioning in a variety of styles. Does seem to have some issues generating the horns occasionally. Also, sometimes you may need to use "dragon" on some models if it doesnt generate properly. I also find that "wings, horns, and tail" help with generation.

Dexdorugoramon SDXL V1

Pony version of alphamon. Trained on both regular alphamon as well as oryuken for about 25 images total. Seems to default to a capeless and weaponless alphamon without his wing pylon thingies, but does work with multiple style/ concepts. Can generate with both his energy sword and oryuken depending on which prompts you use. For his regular sword, use "energy sword,holding sword", for oryuken, use "polearm,holding polearm,wings". Note that since fewer images of oryuken were used, it may take multiple generations to get correctly.

Alphamon PonyXL

Drastically improved version of the Pony model trained at 512X512 for fewer epochs. Its decently flexible and seems to produce consistent results at a strength of 1.0. It does seem to require "wings" and "tail" to get the character to generate completely though, but if your using it to generate random other things just "dorugoramon" would be fine.

Dorugoramon Pony V2.0

Pony version of the dexdorugoramon model. Works good at a variety of weights, though it becomes more flexible at lower weights in exchange for requiring more prompting.

DexDorugoramon Pony

Base 1.5 version. Far less consistent than the XL version. Trained on oryuken but due to 1.5's general inability to generate weapons it cant really generate the polearm. Also has difficulty with the wings in some instances. It can generate his energy sword given enough batches. Otherwise works in both 2d and realistic styles, though I recommend using "mecha,no humans" when using it with realistic models because it likes to make him a person.

Alphamon SD1.5

Dorugoramon SD1.5 V1

Very much in progress version of the model since SD1.5 has difficulty generation dragons and Dexdorugoramon is so complicated, This version is honestly more of a style lora than an actual character one but it can occasionally produce something resembling the subject material. I find that "wings","horns","tail",and,"claws" seem to help with getting the likeness

Alphamon/Dorugoramon/Dexdorugoramon (Digimon)