Ver Anima V2
Trained base on Anima V0.1 base with LoRA_Easy_Training_Scripts. See about this version for training details. Triggerword Added.
Ver Anima V1
Trained base on Anima V0.1 base with LoRA_Easy_Training_Scripts. See about this version for training details.
Ver Chenkin Rectified-Flow V1
Trained base on Chenkin Rectified-Flow 0.3 with LoRA_Easy_Training_Scripts. See about this version for training details.
Ver NOOB V-PRED V4
Trained based on NOOB V-PRED 1.0 with updated dataset and improved parameters.
Ver NOOB V-PRED V3
Trained base on NOOB V-PRED 1.0. See about this version for training details.
After some experiments, I found that the improvements in V2 were actually a manifestation of insufficient training. Bouguereau's paintings do not have such high saturation colors, so I adjusted the training parameters and reduced the minSNR to 1 to maximize the capture of art style. This brought some distortion to the limbs, but I think it is still acceptable.
Ver NOOB V-PRED V2
Trained base on NOOB V-PRED 1.0 with OneTrainer. See about this version for training details. Better in color and prompt responsive responsiveness.
Introduction
A Lycoris to generate Bouguereau Style. Trained based on NOOB V0.65S. Its OUT00-11 Blocks form the foundation of opnmf style fusion. Use photorealistic and oil painting \(medium\) in generation.
Pros
Style generated
Cons
Limb distortion
Blurred eyes
Description
[[subsets]]
caption_dropout_rate = 0.05
caption_extension = ".txt"
caption_tag_dropout_rate = 0.05
image_dir = "G:/dataset/style/oilpainting_V1/5_oilpainting"
keep_tokens = 1
name = "5"
num_repeats = 5
random_crop_padding_percent = 0.05
shuffle_caption = true
[train_mode]
train_mode = "lora"
[general_args.args]
persistent_data_loader_workers = true
vae_batch_size = 5
pretrained_model_name_or_path = ""
mixed_precision = "bf16"
gradient_checkpointing = true
gradient_accumulation_steps = 1
seed = 42
max_data_loader_n_workers = 1
max_token_length = 225
prior_loss_weight = 1.0
sdpa = true
max_train_epochs = 25
cache_latents = true
cache_latents_to_disk = true
[general_args.dataset_args]
resolution = 1024
batch_size = 1
[network_args.args]
network_dim = 64
network_alpha = 64.0
min_timestep = 0
max_timestep = 1000
network_train_unet_only = true
[optimizer_args.args]
optimizer_type = "ProdigyPlusScheduleFree"
lr_scheduler = "constant"
loss_type = "l2"
learning_rate = 1.0
unet_lr = 1.0
max_grad_norm = 1.0
min_snr_gamma = 1.0
[saving_args.args]
output_dir = "G:/LoRA_Easy_Training_Scripts/output"
output_name = "oilpainting-anima1.0-V1-locon-dim64conv16AR1-NewTag-Keep-Real"
save_precision = "bf16"
save_model_as = "safetensors"
save_every_n_epochs = 1
save_toml = true
save_toml_location = "G:/LoRA_Easy_Training_Scripts/output"
[logging_args.args]
log_prefix_mode = "disabled"
run_name_mode = "default"
[anima_args.args]
pretrained_model_name_or_path = "G:/sd-webui-forge-neo/models/Stable-diffusion/ANIMA/anima_baseV10.safetensors"
qwen3 = "G:/sd-webui-forge-neo/models/text_encoder/qwen_3_06b_base.safetensors"
vae = "G:/sd-webui-forge-neo/models/VAE/qwen_image_vae.safetensors"
qwen3_max_token_length = 512
t5_max_token_length = 512
timestep_sampling = "sigmoid"
sigmoid_scale = 1.0
[edm_loss_args.args]
edm2_loss_weighting = false
[bucket_args.dataset_args]
enable_bucket = true
min_bucket_reso = 256
max_bucket_reso = 2048
bucket_reso_steps = 64
[network_args.args.network_args]
conv_dim = 16
conv_alpha = 16.0
train_llm_adapter = "False"
[optimizer_args.args.optimizer_args]
betas = "0.95,0.99"
beta3 = "None"
weight_decay = "0"
weight_decay_by_lr = "True"
d0 = "2e-5"
d_coef = "2"
d_limiter = "True"
schedulefree_c = "0"
eps = "1e-8"
split_groups = "True"
split_groups_mean = "False"
factored = "True"
factored_fp32 = "True"
use_bias_correction = "True"
use_stableadamw = "True"
use_schedulefree = "True"
use_speed = "False"
stochastic_rounding = "True"
fused_back_pass = "False"
use_cautious = "False"
use_grams = "False"
use_adopt = "False"
use_orthograd = "False"
use_focus = "False"
FAQ
Comments (1)
Awesome style!