<h1><strong>v1.0</strong></h1><p>A LoRA of <a target="_blank" rel="ugc" href="https://shinycolors.idolmaster.jp/idol/ikarugaluca/luca.html"><strong><u>Ikaruga Luca</u></strong></a> from THE iDOLM@STER Shiny Colors based on 65 fanarts.</p><h2>Usage</h2><p>The character tag is <code>ikaruga luca</code>.</p><p>Top related tags:</p><pre><code>1girl, solo, black hair, blonde hair, multicolored hair, yellow eyes, looking at viewer, piercing, streaked hair, ear piercing, breasts, simple background, jewelry, choker, jacket, collarbone, off shoulder, crop top, shirt, black jacket, medium hair, necklace, white background, bare shoulders, navel, midriff, nail polish, upper body, blush, two-tone hair, holding, fishnets, off-shoulder shirt, long sleeves, medium breasts, skirt, cleavage, small breasts, black choker, dress
</code></pre><p>For specific outfits, refer to the preview images or the dataset.</p><h2>Training info</h2><p>Dataset: <a target="_blank" rel="ugc" href="https://huggingface.co/datasets/gustproof/sd-data/blob/main/luca.zip"><strong><u>65 fanarts from Danbooru</u></strong></a></p><p>Training cost: ~0.5 T4-hour</p><p>Training config:</p><pre><code>[model_arguments]
v2 = false
v_parameterization = false
pretrained_model_name_or_path = "animefull-final-pruned.ckpt"

[additional_network_arguments]
no_metadata = false
unet_lr = 0.0001
text_encoder_lr = 0.0001
network_module = "networks.lora"
network_dim = 1
network_alpha = 1
network_train_unet_only = false
network_train_text_encoder_only = false

[optimizer_arguments]
min_snr_gamma = 5
optimizer_type = "AdamW8bit"
learning_rate = 0.0001
max_grad_norm = 1.0
lr_scheduler = "constant"
lr_warmup_steps = 0

[dataset_arguments]
dataset_repeats = 1
shuffle_caption = true
keep_tokens = 1
caption_dropout_rate = 0
caption_tag_dropout_rate = 0
caption_dropout_every_n_epochs = 0
color_aug = false
token_warmup_min = 1
token_warmup_step = 0

[training_arguments]
save_precision = "fp16"
save_every_n_epochs = 1
train_batch_size = 4
max_token_length = 225
mem_eff_attn = false
xformers = true
max_train_epochs = 9999
max_data_loader_n_workers = 8
persistent_data_loader_workers = true
gradient_checkpointing = false
gradient_accumulation_steps = 1
mixed_precision = "fp16"
clip_skip = 2
lowram = true

[saving_arguments]
save_model_as = "safetensors"</code></pre>

Files with SHA256: 50026467b2970f069dcd8a0aca4316e3e3e5f59f4680bd520f63b353d2230908