{
  "_class_name": "UNet2DConditionModel",
  "_diffusers_version": "0.2.4",
  "act_fn": "silu",
  "attention_head_dim": 8,
  "block_out_channels": [
    128,
    128,
    256,
    256,
    512,
    512
  ],
  "center_input_sample": false,
  "create_text_encoder": false,
  "cross_attention_dim": 768,
  "down_block_types": [
    "DownBlock2D",
    "DownBlock2D",
    "CrossAttnDecoderPositionEncoderPositionDownBlock2D",
    "CrossAttnDecoderPositionEncoderPositionDownBlock2D",
    "CrossAttnDecoderPositionEncoderPositionDownBlock2D",
    "CrossAttnDecoderPositionEncoderPositionDownBlock2D"
  ],
  "downsample_padding": 1,
  "flip_sin_to_cos": true,
  "freq_shift": 0,
  "gradient_checkpointing": false,
  "in_channels": 1,
  "layers_per_block": 2,
  "max_l": 1025,
  "mid_block_scale_factor": 1,
  "mid_block_type": "UNetMidBlock2DCrossAttnDecoderPositionEncoderPosition",
  "norm_eps": 1e-05,
  "norm_num_groups": 32,
  "out_channels": 1,
  "sample_size": [
    64,
    320
  ],
  "up_block_types": [
    "CrossAttnDecoderPositionEncoderPositionUpBlock2D",
    "CrossAttnDecoderPositionEncoderPositionUpBlock2D",
    "CrossAttnDecoderPositionEncoderPositionUpBlock2D",
    "CrossAttnDecoderPositionEncoderPositionUpBlock2D",
    "UpBlock2D",
    "UpBlock2D"
  ]
}