{
  "_class_name": "CVVAESD3Model",
  "_diffusers_version": "0.29.2",
  "act_fn": "silu",
  "block_out_channels": [
    128,
    256,
    512,
    512
  ],
  "causal_decoder": false,
  "causal_encoder": true,
  "double_z": true,
  "down_block_types": [
    "DownEncoderBlock3D",
    "DownEncoderBlock3D",
    "DownEncoderBlock3D",
    "DownEncoderBlock3D"
  ],
  "en_de_n_frames_a_time": 16,
  "half_3d": true,
  "in_channels": 3,
  "layers_per_block": 2,
  "mid_block_add_attention": true,
  "norm_num_groups": 32,
  "num_video_frames": null,
  "out_channels": 16,
  "reshape_x_dim_to_4": false,
  "reshape_z_dim_to_4": false,
  "spatial_n_compress": 8,
  "tile_overlap_ratio": 0.2222,
  "tile_spatial_size": 576,
  "time_n_compress": 4,
  "up_block_types": [
    "UpDecoderBlock3D",
    "UpDecoderBlock3D",
    "UpDecoderBlock3D",
    "UpDecoderBlock3D"
  ]
}