Experimental, preliminary models.

I would appreciate any comments on how they work!

Some models may be deleted after a certain period of time.

ZImageTurbo

• NOHVEIL v0.2 - ...New

• NOHVEIL v0.1 -

Chroma v1.0HD

When creating images of this model, it is recommended to use a resolution of 1024x1024.

• Chrgai v0.1 - I was sure that the model had completely failed to convey the style, but despite this, it performed much better with negative hints.

• ugly face, watermark, mutated hands, bad eyes, bad teeth, bad arms, missing limbs, poorly drawn face, three feet, fused feet, fused thigh, three thigh, fused thigh, extra thigh, missing fingers, extra fingers, ugly fingers, amputation, square jawline, mutilated, wide hips, deformed, disconnected limbs, worst quality, low quality, jpeg artifacts

Pony 6

score_9, score_8_up, score_7_up, score_6_up,

• Omega Sentinel v0.1 - When developing this set, I overlooked one important detail: I didn't remove the background from the character. In my opinion, this could have certain consequences.

Styl:

• Incase T v1.0

• Incase F v1.0 - As a result of training on 20 images, it was found that the best performance of the model is achieved when using a coefficient of 0.8, which minimizes distortion.

• Hoshino_Akari v0.1 - I didn't choose the most successful style for the test. The file size is 13 Mb🙀. This model allows you to draw details in more detail.

• p6Keyframe v1.0 - One of the failed models. Si desea lograr un efecto similar, no use colores como el pelo púrpura, la camisa roja, etc. en la descripción.

• GAIXL v1.1 - Previously, I used the original image resolution, but now I've switched to training with a resolution of ???? x 1024. I believe this is more effective for conveying style in Pony6.

WAI-NSFW-illustrious-SDXL v15

Styl:

• Incase F v0.2 - More steps in learning.

• Incase F v0.1 - Unfortunately, the training is not in the pairs I wanted, but as they say, what happened, happened. If it weren't for your positive feedback, I wouldn't have continued working on this model.

Illustrious v0.1

Me parece que este modelo es el más adecuado para aprender el estilo de anime.

Styl:

• GAIIL v0.1 - I don't think the Illustrious v0.1 model captures the style as well as the Pony v6 and Flux.1D models (although I could be wrong).

• INCFIL v0.1 - I don't know who likes the Illustrious v0.1 model, but I don't really like it. Maybe I just haven't had enough time to work with it.

Flux.1D

Styl:

• ACMS1 v1.0 - Continues experimenting with Flux. This model was trained using examples of Asian women.

• INCFF1 v0.1 - I like Flux.1D, but I'm annoyed by its censorship policy and reluctance to use keywords, so maybe I need a different approach. I'm also concerned about the size of the models. However, I've noticed that Flux has a better understanding of what I want to display.