Read the contents below before use.

• Trigger word and usage guide are listed below.

• All sample images were generated with hire.fx, with an upscaling value of about 1.3x.

ABOUT THE MODEL

Miiyamoto Hikari from SUMMER LESSON

20250902v0.2

Trained on Chroma v1.0 HD

Reduced the model’s size, added a new set of clothing, and increased the training steps to improve the probability of correctly learning the clothing concepts in the dataset.

Features

• 4+ outfits

TRIGGER WORDS

All trigger words are based on full body, you can modify them as needed.

Example

You could find the example prompts from my showcase.

My prompts are basically composed in the order of [character traits] + [clothing] + [customize content] + [background], and you can delete or modify them as needed.

School Uniform

miiyamotohikari, a photo of miiyamotohikari, a woman, with black hair, ponytail, bangs, pink scrunchie, brown eyes, she is wearing white collar shirt, blue plaid skirt, blue plaid necktie, red wristband, black kneehighs, bandaid on knee, brown loafers, 

Cyan Tank Top

miiyamotohikari, a photo of miiyamotohikari, a woman, with black hair, bangs, ponytail, pink scrunchie, she is wearing cyan tank top, black shorts,

White Dress

miiyamotohikari, a photo of miiyamotohikari, a woman, with black hair, bangs, pink scrunchie, brown eyes, she is wearing sleeveless white dress, white wristwatch, white flat shoes, 

Swimming Suit

miiyamotohikari, a photo of miiyamotohikari, a woman, with black hair, ponytail, bangs, blue scrunchie, brown eyes, she is blue bikini with white ruffle trim and white polka dots, white ribbon, white flip-flops,

USAGE GUIDE

Requirement

Checkpoints: Chroma v1.0 HD

Recommended Generation Tools: ComfyUI, Forge

Recommended size: 768*1344, 832*1216, 896*1156, 1024*1024, 1344*768, 1216*8

32, 1156*896

I’ve set the gallery moderation of this post to “R”, so explicit content may not appear on this page

ARCHIVES

20250902v0.1

Trained on Chroma v1.0 HD

Chroma is truly impressive — it performs exceptionally well when training 3D characters and realistic humans. At 1024×1024 resolution, you barely need to rely on Adetailer or hire.fx to produce solid results.

All sample images were generated directly at size 768*1344. Using Hire.fx or Adetailer for refinement would achieve even better results.

The current issue is that with my parameters, Chroma often learns facial features at an astonishing speed, while the concept of clothing details remains in an uncooked state.
On the other hand, when using the older booru-style tag captions, Chroma still responds, but not as effectively as with fixed natural language sentences.

Therefore, captions for this model use a combination of tags + natural language.
Also, just like Flux, using certain tags such as white background, black background, xxx background tends to produce blurry images.

Features

• 3+ outfits