config/class_level/wolf/wolf.yaml

pretrained_model_path: "./ckpt/stable-diffusion-v1-5"
logdir: ./result/class_level/wolf2pig_preserve_bg

dataset_config:
    path: "data/wolf/wolf"  
    prompt: A wolf
    start_sample_frame: 0
    n_sample_frame: 16
    sampling_rate: 2
    layout_mask_dir: "data/wolf/layout_masks/"
    layout_mask_order: ['wolf','bg'] 
    negative_promot: "a wolf, deformed"

control_config:
    control_type: "depth_zoe"
    pretrained_controlnet_path: "./ckpt/sd-controlnet-depth"
    controlnet_conditioning_scale: 0.7


editing_config:
    use_invertion_latents: true
    inject_step: 0
    old_qk: 1
    flatten_res: [1,2,4]
    guidance_scale: 7.5
    use_pnp: false
    use_freeu: false
    editing_prompts: [     
            ['A cute pig in the autum forest','A cute pig','autum forest'],
            ['A husky in the green forest','A husky','green forest'],
            ['A tiger in the autum forest','A tiger','autum forest'],
            ['A bear in the autum  forest','A bear','autum forest'], 

    ]
    clip_length: "${..dataset_config.n_sample_frame}"
    sample_seeds: [0]
    num_inference_steps: 50
    blending_percentage: 0
    modulated_percentage: 0.3

test_pipeline_config:
    target: video_diffusion.pipelines.ddim_spatial_temporal.DDIMSpatioTemporalStableDiffusionPipeline
    num_inference_steps: "${..validation_sample_logger.num_inference_steps}"



seed: 42