{
  "_class_name": "WFVAEModel",
  "_diffusers_version": "0.30.2",
  "attention_type": "AttnBlock3DFix",
  "base_channels": 192,
  "connect_res_layer_num": 1,
  "decoder_energy_flow_hidden_size": 128,
  "decoder_num_resblocks": 2,
  "dropout": 0.0,
  "encoder_energy_flow_hidden_size": 128,
  "encoder_num_resblocks": 2,
  "l1_dowmsample_block": "Downsample",
  "l1_downsample_wavelet": "HaarWaveletTransform2D",
  "l1_upsample_block": "Upsample",
  "l1_upsample_wavelet": "InverseHaarWaveletTransform2D",
  "l2_dowmsample_block": "Spatial2xTime2x3DDownsample",
  "l2_downsample_wavelet": "HaarWaveletTransform3D",
  "l2_upsample_block": "Spatial2xTime2x3DUpsample",
  "l2_upsample_wavelet": "InverseHaarWaveletTransform3D",
  "latent_dim": 16,
  "norm_type": "layernorm",
  "scale": [
    0.18215,
    0.18215,
    0.18215,
    0.18215
  ],
  "shift": [
    0,
    0,
    0,
    0
  ],
  "t_interpolation": "trilinear",
  "use_attention": true
}