unet/config.json

{
  "_class_name": "UNet3DConditionModel",
  "_diffusers_version": "0.30.0.dev0",
  "_name_or_path": "/root/.cache/huggingface/hub/models--damo-vilab--text-to-video-ms-1.7b/snapshots/8227dddca75a8561bf858d604cc5dae52b954d01/unet",
  "act_fn": "silu",
  "attention_head_dim": 64,
  "block_out_channels": [
    320,
    640,
    1280,
    1280
  ],
  "cross_attention_dim": 1024,
  "down_block_types": [
    "CrossAttnDownBlock3D",
    "CrossAttnDownBlock3D",
    "CrossAttnDownBlock3D",
    "DownBlock3D"
  ],
  "downsample_padding": 1,
  "in_channels": 4,
  "layers_per_block": 2,
  "mid_block_scale_factor": 1,
  "norm_eps": 1e-05,
  "norm_num_groups": 32,
  "num_attention_heads": null,
  "out_channels": 4,
  "sample_size": 32,
  "time_cond_proj_dim": null,
  "up_block_types": [
    "UpBlock3D",
    "CrossAttnUpBlock3D",
    "CrossAttnUpBlock3D",
    "CrossAttnUpBlock3D"
  ]
}