{
    "_class_name": "GroundedUNet3DConditionModel",
    "_diffusers_version": "0.15.0.dev0",
    "_name_or_path": "/home/tony/text-to-video-lvd-ms-1.7b/unet",
    "act_fn": "silu",
    "attention_head_dim": 64,
    "block_out_channels": [
      320,
      640,
      1280,
      1280
    ],
    "cross_attention_dim": 1024,
    "attention_type": "gated",
    "down_block_types": [
      "CrossAttnDownBlock3D",
      "CrossAttnDownBlock3D",
      "CrossAttnDownBlock3D",
      "DownBlock3D"
    ],
    "in_channels": 4,
    "layers_per_block": 2,
    "norm_eps": 1e-05,
    "norm_num_groups": 32,
    "out_channels": 4,
    "sample_size": 32,
    "up_block_types": [
      "UpBlock3D",
      "CrossAttnUpBlock3D",
      "CrossAttnUpBlock3D",
      "CrossAttnUpBlock3D"
    ]
  }