{
  "variant": "pruna",
  "source": "PrunaAI/PrunaVAED",
  "latent_channels": 128,
  "out_channels": 3,
  "patch_size": 4,
  "patch_size_t": 1,
  "decoder_block_out_channels": [
    128,
    256,
    384,
    1024
  ],
  "decoder_layers_per_block": [
    4,
    6,
    4,
    2,
    2
  ],
  "decoder_spatio_temporal_scaling": [
    true,
    true,
    true,
    true
  ],
  "decoder_inject_noise": [
    false,
    false,
    false,
    false,
    false
  ],
  "decoder_spatial_padding_mode": "zeros",
  "decoder_causal": false,
  "upsample_factor": [
    2,
    2,
    1,
    2
  ],
  "upsample_residual": [
    false,
    false,
    false,
    false
  ],
  "upsample_type": [
    "spatiotemporal",
    "spatiotemporal",
    "temporal",
    "spatial"
  ],
  "timestep_conditioning": false,
  "resnet_norm_eps": 1e-06,
  "scaling_factor": 1.0
}