All Images Here Are Generated By AI.

All Models

>

THE iDOLM@STER Shiny Colors - Sprites (style, 20+ chars)

THE iDOLM@STER Shiny Colors - Sprites (style, 20+ chars)

Generate AI Images. Simple and Easy

AIEasyPic is the #1 AI image generator. Generate images with simple prompts, swap faces, train models to take AI photos of yourself, and more.

Model Description by Creator

A LoRA based on 1186 Shiny Colors character sprites, recognizing 20+ characters to varying degrees. This is a search for answers to the following questions:

  • Does the lack of data variety simplify fitting?

  • Can the model generalize from data lacking variety?

The answers to both seem to be "not really".

Note: the site does not display txt2img parameters with non-ASCII characters. To see how the preview images were generated, download and inspect them locally.

Cover image params:

八宮めぐる, 1girl, blonde hair, ahoge, solo, smile, long hair, outdoors, sunset, beach, blue eyes, jewelry, looking at viewer, braid, necklace, jacket, ocean, holding, sky, cloud, collarbone, tank top, breasts, white tank top, bangs, sand
<lora:shiny_colors_sprites:1>
Negative prompt: worst quality, out of frame, abstract background
Steps: 24, Sampler: DPM++ 2M Karras, CFG scale: 5.5, Seed: 2652041511, Size: 512x768, Model hash: c8c497baf5, Model: anything-v3-1

Usage

The style itself doesn't have a trigger word. For the characters, below are the character tags and respective top 20 related tags.

櫻木真乃,
1girl, solo, brown eyes, looking at viewer, smile, simple background, white background, blush, brown hair, light brown hair, own hands together, bangs, skirt, hair ornament, frills, shirt, medium hair, collarbone, short sleeves, long sleeves

風野灯織,
1girl, solo, black hair, white background, simple background, looking at viewer, mole under mouth, mole, bangs, long hair, hand on hip, closed mouth, ponytail, black eyes, skirt, long sleeves, collarbone, jewelry, hair ornament, shirt

八宮めぐる,
1girl, solo, blonde hair, blue eyes, smile, looking at viewer, white background, simple background, long hair, ahoge, breasts, hair ornament, bangs, twintails, hairclip, low twintails, closed mouth, medium breasts, skirt, shirt

月岡恋鐘,
1girl, solo, brown hair, smile, looking at viewer, white background, blush, simple background, long hair, breasts, brown eyes, bangs, large breasts, closed mouth, dress, ribbon, cleavage, jewelry, frills, skirt

西城樹里,
1girl, solo, blonde hair, simple background, looking at viewer, purple eyes, short hair, white background, bangs, shirt, collarbone, hair between eyes, shorts, jewelry, jacket, long sleeves, skirt, standing, hair ornament, white shirt

杜野凛世,
1girl, solo, white background, red eyes, simple background, looking at viewer, bangs, holding, closed mouth, black hair, hair between eyes, blue hair, doll, standing, short hair, hair ornament, holding doll, long sleeves, hair bun, dress

大崎甜花,
1girl, solo, simple background, looking at viewer, blush, white background, brown hair, yellow eyes, bangs, long hair, hair between eyes, dress, long sleeves, hair ornament, frills, own hands together, bow, hands up, standing, feet out of frame

田中摩美々,
1girl, solo, purple hair, bangs, purple eyes, looking at viewer, simple background, white background, diagonal bangs, jewelry, earrings, twintails, breasts, choker, collarbone, bare shoulders, thighhighs, nail polish, long sleeves, skirt

小宮果穂,
1girl, solo, smile, looking at viewer, red hair, white background, red eyes, simple background, bangs, long hair, blush, hair between eyes, ahoge, shorts, long sleeves, shirt, collarbone, jacket, skirt, closed mouth

大崎甘奈,
1girl, solo, looking at viewer, white background, brown hair, smile, simple background, yellow eyes, bangs, long hair, closed mouth, collarbone, breasts, dress, frills, jewelry, bare shoulders, medium breasts, hair ornament, skirt

幽谷霧子,
1girl, solo, grey hair, purple eyes, smile, looking at viewer, white background, simple background, long hair, blush, bangs, twintails, closed mouth, dress, breasts, bandaid, collarbone, bandages, frills, bare shoulders

園田智代子,
1girl, solo, brown hair, smile, white background, looking at viewer, simple background, red eyes, bangs, twintails, hair bun, double bun, skirt, breasts, hair ornament, long sleeves, jewelry, short twintails, blush, brown eyes

三峰結華,
1girl, solo, blue eyes, black hair, looking at viewer, smile, white background, simple background, hand on hip, long hair, closed mouth, bangs, twintails, glasses, skirt, collarbone, long sleeves, hat, dress, breasts

有栖川夏葉,
1girl, solo, hand on hip, smile, looking at viewer, white background, simple background, bangs, orange hair, long hair, black eyes, breasts, jewelry, earrings, red hair, medium breasts, collarbone, cleavage, standing, skirt

桑山千雪,
1girl, solo, brown hair, looking at viewer, white background, smile, blush, simple background, brown eyes, ahoge, breasts, long hair, bangs, braid, large breasts, dress, single braid, jewelry, braided ponytail, hair over shoulder

芹沢あさひ,
1girl, solo, white background, smile, blue eyes, looking at viewer, simple background, short hair, bangs, grey hair, closed mouth, shorts, long sleeves, jewelry, earrings, gloves, skirt, jacket, shirt, feet out of frame

白瀬咲耶,
1girl, solo, black hair, looking at viewer, yellow eyes, hand on hip, white background, bangs, simple background, hair between eyes, breasts, long hair, smile, ponytail, jewelry, cleavage, collarbone, large breasts, shirt, earrings

黛冬優子,
1girl, solo, long hair, black hair, bangs, looking at viewer, white background, smile, simple background, brown eyes, blunt bangs, blush, skirt, two side up, closed mouth, breasts, ribbon, long sleeves, frills, standing

和泉愛依,
1girl, solo, brown hair, blonde hair, multicolored hair, gradient hair, white background, simple background, black eyes, long hair, smile, looking at viewer, breasts, dark skin, earrings, jewelry, dark-skinned female, gyaru, tan, hand on hip

浅倉透,
1girl, solo, simple background, white background, short hair, looking at viewer, multicolored hair, smile, bangs, jewelry, earrings, blue hair, gradient hair, ahoge, blue eyes, brown hair, parted bangs, nail polish, collarbone, shirt

樋口円香,
1girl, solo, mole under eye, mole, white background, looking at viewer, simple background, short hair, brown hair, purple eyes, bangs, hairclip, hair ornament, nail polish, skirt, jacket, jewelry, shirt, long sleeves, collarbone

市川雛菜,
1girl, solo, white background, brown hair, blush, simple background, smile, looking at viewer, brown eyes, bangs, parted bangs, long hair, hair bun, skirt, ribbon, nail polish, hair ribbon, short sleeves, shirt, dress

七草にちか,
1girl, solo, looking at viewer, smile, simple background, white background, green hair, green eyes, short hair, blush, skirt, short sleeves, shirt, collarbone, hair ornament, jewelry, jacket, shorts, bangs, light smile

緋田美琴,
1girl, solo, multicolored hair, brown hair, looking at viewer, simple background, white background, blonde hair, gradient hair, multicolored eyes, jewelry, jacket, earrings, necklace, dress, nail polish, breasts, collarbone, light smile, navel

福丸小糸,
1girl, solo, black hair, looking at viewer, white background, purple eyes, twintails, simple background, bangs, hair ornament, skirt, ribbon, hair ribbon, long sleeves, dress, smile, gloves, shirt, short sleeves, pleated skirt

七草はづき,
1girl, solo, green hair, looking at viewer, white background, simple background, smile, green eyes, open mouth, index finger raised, blush, shirt, hand up, bangs, long hair, breasts, skirt, pantyhose, collarbone, :d

斑鳩ルカ,
1girl, solo, blonde hair, multicolored hair, white background, black hair, looking at viewer, simple background, yellow eyes, jewelry, streaked hair, medium hair, piercing, two-tone hair, bangs, ear piercing, belt, breasts, standing, leopard print

As the training data is fairly homogeneous, the use of negative prompts and emphases may be required.

Training info


Trained solely on 1186 Shiny Colors character sprites extracted from ShinyColorsDB-SpineViewer without balancing and tag cleaning. Sprites of the same character were extracted in the same pose. Exact duplicates were merged. Sprites were padded with a white background to at least 500px in width. Tags generated with SmilingWolf/wd-v1-4-convnextv2-tagger-v2 with a threshold of 0.35. Caption template was "[char name], [other tags shuffled]".

Training cost: ~2 T4-hours

Data distribution:

       character  total
1       櫻木真乃     59
2       風野灯織     56
3      八宮めぐる    55
4       月岡恋鐘     55
5       西城樹里     54
6       杜野凛世     54
7       大崎甜花     54
8      田中摩美々    53
9       小宮果穂     53
10      大崎甘奈     53
11      幽谷霧子     52
12     園田智代子    52
13      三峰結華     51
14     有栖川夏葉    51
15      桑山千雪     51
16     芹沢あさひ    50
17      白瀬咲耶     48
18      黛冬優子     44
19      和泉愛依     43
20       浅倉透      35
21      樋口円香     34
22      市川雛菜     32
23     七草にちか    32
24      緋田美琴     29
25      福丸小糸     28
26     七草はづき     6
27      斑鳩ルカ      2

LoRA meta dump:

{'ss_cache_latents': 'False',
 'ss_caption_dropout_every_n_epochs': '0',
 'ss_caption_dropout_rate': '0.0',
 'ss_caption_tag_dropout_rate': '0.0',
 'ss_clip_skip': '2',
 'ss_dataset_dirs': '{"meta_lat.json": {"n_repeats": 1, "img_count": 1186}}',
 'ss_datasets': '[{"is_dreambooth": false, "batch_size_per_device": 4, "num_train_images": 1186, "num_reg_images": 0, "resolution": [640, 640], "enable_bucket": true, "min_bucket_reso": null, "max_bucket_reso": null, "tag_frequency": {"meta_lat.json": {}}, "bucket_info": {"buckets": {"0": {"resolution": [384, 768], "count": 71}, "1": {"resolution": [384, 832], "count": 3}, "2": {"resolution": [384, 896], "count": 2}, "3": {"resolution": [384, 960], "count": 1}, "4": {"resolution": [448, 704], "count": 58}, "5": {"resolution": [448, 768], "count": 172}, "6": {"resolution": [448, 832], "count": 419}, "7": {"resolution": [448, 896], "count": 183}, "8": {"resolution": [512, 640], "count": 6}, "9": {"resolution": [512, 704], "count": 60}, "10": {"resolution": [512, 768], "count": 201}, "11": {"resolution": [576, 576], "count": 2}, "12": {"resolution": [576, 640], "count": 4}, "13": {"resolution": [576, 704], "count": 3}, "14": {"resolution": [704, 512], "count": 1}}, "mean_img_ar_error": 0.0}, "subsets": [{"img_count": 1186, "num_repeats": 1, "color_aug": false, "flip_aug": false, "random_crop": false, "shuffle_caption": true, "keep_tokens": 1, "image_dir": "train_data", "metadata_file": "meta_lat.json"}]}]',
 'ss_epoch': '10',
 'ss_face_crop_aug_range': 'None',
 'ss_full_fp16': 'False',
 'ss_gradient_accumulation_steps': '1',
 'ss_gradient_checkpointing': 'False',
 'ss_learning_rate': '1',
 'ss_lowram': 'True',
 'ss_lr_scheduler': 'polynomial',
 'ss_lr_warmup_steps': '300',
 'ss_max_grad_norm': '1.0',
 'ss_max_token_length': '225',
 'ss_max_train_steps': '3020',
 'ss_mixed_precision': 'fp16',
 'ss_network_alpha': '32',
 'ss_network_dim': '32',
 'ss_network_module': 'lycoris.kohya',
 'ss_new_sd_model_hash': '89d59c3dde4c56c6d5c41da34cc55ce479d93b4007046980934b14db71bdb2a8',
 'ss_noise_offset': 'None',
 'ss_num_batches_per_epoch': '302',
 'ss_num_epochs': '10',
 'ss_num_reg_images': '0',
 'ss_num_train_images': '1186',
 'ss_optimizer': 'dadaptation.dadapt_adam.DAdaptAdam',
 'ss_output_name': 'shiny_colors_sprites',
 'ss_prior_loss_weight': '1.0',
 'ss_sd_model_hash': '925997e9',
 'ss_sd_model_name': 'Animefull-final-pruned.ckpt',
 'ss_sd_scripts_commit_hash': '7baab00cec018077e3550b5d358f51f713618c3d',
 'ss_seed': 'None',
 'ss_session_id': '3748380245',
 'ss_tag_frequency': '{"meta_lat.json": {}}',
 'ss_text_encoder_lr': '1.0',
 'ss_training_comment': 'None',
 'ss_training_finished_at': '1679406477.8537939',
 'ss_training_started_at': '1679399317.376039',
 'ss_unet_lr': '1.0',
 'ss_v2': 'False',
 'sshs_legacy_hash': '497ad236',
 'sshs_model_hash': '42ffb28c196e8ca08a5831730250cc1e47637bb18be7731f1f81ac18bdb8ca2b'}

Try Full Demo
Quick Generate
Image quality might not be as good as what you can achieve with our full generation tool.
Placeholder
Choose A Style, Type What You Want In The Image And Click Generate
Placeholder
Choose A Style, Type What You Want In The Image And Click Generate
What you want in the image.
Clear
Random
* This model is available for use and will generate similar character to the ones seen in the images.
Train Your Own Model

Images Generated With This Model

Loading...

Generate AI Images. Simple and Easy

AIEasyPic is the #1 AI image generator. Generate images with simple prompts, swap faces, train models to reimagine yourself, and more.