I was trying to train my Au Ra in FF14, by providing screenshot from the game. Its can output the horn and scales right, but the 3D style is affecting the generation. Anyone have a idea of how to fix it?These are some of the settingpretrained_model_name_or_path = "./sd-models/latest.ckpt"resolution = "640,640"enable_bucket = truemin_bucket_reso = 256max_bucket_reso = 1024save_every_n_epochs = 1max_train_epochs = 10train_batch_size = 2network_train_unet_only = falsenetwork_train_text_encoder_only = falselearning_rate = 0.0001unet_lr = 3.5e-5text_encoder_lr = 3e-6lr_scheduler = "cosine_with_restarts"optimizer_type = "Lion"lr_scheduler_num_cycles = 1network_module = "networks.lora"network_dim = 32network_alpha = 32shuffle_caption = trueweighted_captions = falsekeep_tokens = 1max_token_length = 255seed = 1337prior_loss_weight = 1clip_skip = 2gradient_checkpointing = falsemixed_precision = "fp16"save_precision = "fp16"xformers = truelowram = falsecache_latents = truecache_latents_to_disk = falsepersistent_data_loader_workers = true

Aura (FFxiv)

Merge with Lora:[style_gomzi]with 0.8 : 0.2 ratio. lower the 3D feel of the lora and make more anime output.

Files with SHA256: 75ad968b1d121d769794954bc6d965b1b1ba28a86e992c4ae4b3032ede16d2fa