config.json

{
  "beta_end": 0.02,
  "beta_schedule": "squaredcos_cap_v2",
  "beta_start": 0.0001,
  "clip_sample": true,
  "clip_sample_range": 1.0,
  "crop_is_random": true,
  "crop_shape": [
    432,
    576
  ],
  "diffusion_step_embed_dim": 128,
  "do_mask_loss_for_padding": false,
  "down_dims": [
    512,
    1024,
    2048
  ],
  "horizon": 16,
  "input_normalization_modes": {
    "observation.images.phone": "mean_std",
    "observation.state": "mean_std"
  },
  "input_shapes": {
    "observation.images.phone": [
      3,
      480,
      640
    ],
    "observation.state": [
      6
    ]
  },
  "kernel_size": 5,
  "n_action_steps": 8,
  "n_groups": 8,
  "n_obs_steps": 1,
  "noise_scheduler_type": "DDPM",
  "num_inference_steps": null,
  "num_train_timesteps": 100,
  "output_normalization_modes": {
    "action": "mean_std"
  },
  "output_shapes": {
    "action": [
      6
    ]
  },
  "prediction_type": "sample",
  "pretrained_backbone_weights": "ResNet18_Weights.IMAGENET1K_V1",
  "spatial_softmax_num_keypoints": 32,
  "use_film_scale_modulation": true,
  "use_group_norm": false,
  "vision_backbone": "resnet18"
}