{
  "_class_name": "UNetSpatioTemporalConditionModel",
  "_diffusers_version": "0.24.0",
  "addition_time_embed_dim": 1,
  "block_out_channels": [
    128,
    256,
    256,
    512
  ],
  "cross_attention_dim": 1,
  "decay": 0.9999,
  "down_block_types": [
    "CrossAttnDownBlockSpatioTemporal",
    "CrossAttnDownBlockSpatioTemporal",
    "CrossAttnDownBlockSpatioTemporal",
    "DownBlockSpatioTemporal"
  ],
  "in_channels": 8,
  "inv_gamma": 1.0,
  "layers_per_block": 2,
  "min_decay": 0.0,
  "num_attention_heads": [
    8,
    16,
    16,
    32
  ],
  "num_frames": 64,
  "optimization_step": 80000,
  "out_channels": 4,
  "power": 0.6666666666666666,
  "projection_class_embeddings_input_dim": 1,
  "sample_size": 14,
  "transformer_layers_per_block": 1,
  "up_block_types": [
    "UpBlockSpatioTemporal",
    "CrossAttnUpBlockSpatioTemporal",
    "CrossAttnUpBlockSpatioTemporal",
    "CrossAttnUpBlockSpatioTemporal"
  ],
  "update_after_step": 0,
  "use_ema_warmup": false
}