config: name: flux_dev_Indian_anime_fusion_v1 process: - datasets: - cache_latents_to_disk: true caption_dropout_rate: 0.2 caption_ext: txt folder_path: /root/lorahub/flux_dev_Indian_anime_fusion_v1/dataset resolution: - 512 - 768 - 1024 shuffle_tokens: false token_dropout_rate: 0.01 device: cuda:0 model: is_flux: true name_or_path: black-forest-labs/FLUX.1-dev quantize: true text_encoder_bits: 8 network: linear: 42 linear_alpha: 42 transformer_only: true type: lora performance_log_every: 500 sample: height: 1024 neg: '' prompts: - a woman selling ice cream in dairy farm [trigger] - a woman sitting on sofa and taking care of her pet [trigger] - a man doing farming in hot summer with bull [trigger] - worshiping the cow is the purity in Hinduism [trigger] - children celebrating diwali in light, crackers and lightning [trigger] - a businessman got first prize in his field, smiling and thanking on stage [trigger] sample_every: 500 sample_steps: 25 sampler: flowmatch seed: 42 walk_seed: true width: 1024 save: dtype: float16 max_step_saves_to_keep: 3 save_every: 500 save_format: diffusers train: batch_size: 1 dtype: bf16 ema_config: ema_decay: 0.99 use_ema: true gradient_accumulation_steps: 1 gradient_checkpointing: true linear_timesteps: true loss_type: mse lr: 0.0002 noise_scheduler: flowmatch optimizer: adamw8bit reg_weight: 1.0 steps: 3000 target_noise_multiplier: 1.0 train_text_encoder: false train_unet: true training_folder: /root/lorahub trigger_word: ' indo_swaping_chan ' type: sd_trainer job: extension meta: description: This LoRa model is designed to capture the fusion of Indian cultural elements with anime aesthetics. It emphasizes intricate details in traditional attire, jewelry, and patterns, while maintaining the expressive, stylized look of anime art. The result is a unique blend that highlights Indian heritage through a colorful, high-quality anime style with depth and character.