---
license: apache-2.0
base_model:
- circlestone-labs/Anima
tags:
- peft
- lora
- adapter
---

# Experimental Anima LLLite Regional Controlnet

Apply Anima ControlNet-LLLite parameters:
![image](https://cdn-uploads.huggingface.co/production/uploads/68794279700df7682fb2a81f/R8nNSAsGsJoYWhAI8IvjQ.png)

Control image example: (Use basic colors as region)
![rgthree.compare._temp_fobmy_00002_](https://cdn-uploads.huggingface.co/production/uploads/68794279700df7682fb2a81f/ahsOihVAw8QLXDtwCwRzf.png)

Prompt: (kinda works)  
  base prompt  
  region color: region prompt  

## Training Parameters:

caption_extension: .txt  
shuffle_caption: false  
resolution: 1024  
batch_size: 1  
enable_bucket: true  
bucket_no_upscale: true  
bucket_reso_steps: 16  
min_bucket_reso: 64  
max_bucket_reso: 1536  
num_repeats: 2  

580 image/control/caption pairs  
580 * 2 repeats = 1160 steps per epoch  
8 epochs = 9280 total steps  

learning_rate: 3e-4  
max_train_epochs: 8  
seed: 42  
optimizer: AdamW8bit  
lr_scheduler: constant  
mixed_precision: bf16  
save_precision: bf16  
save_model_as: safetensors  
save_every_n_epochs: 1  
gradient_checkpointing: enabled  
cache_latents_to_disk: enabled  
cache_text_encoder_outputs_to_disk: enabled  
vae_chunk_size: 64  
vae_disable_cache: enabled  

timestep_sampling: shift  
discrete_flow_shift: 3.0  
attn_mode: torch  

cond_emb_dim: 32  
lllite_cond_dim: 64  
lllite_mlp_dim: 64  
lllite_target_layers: self_attn_qkv  
lllite_cond_resblocks: 4  
lllite_use_aspp: enabled  
caption_dropout_rate: 0.15