transformer/config.json

{
  "_class_name": "WanTransformerAudioMask3DModel",
  "_diffusers_version": "0.32.2",
  "add_control_adapter": false,
  "add_ref_conv": false,
  "cross_attn_norm": true,
  "dim": 1536,
  "eps": 1e-06,
  "ffn_dim": 8960,
  "freq_dim": 256,
  "hidden_size": 1536,
  "in_channels": 36,
  "in_dim": 36,
  "in_dim_control_adapter": 24,
  "in_dim_ref_conv": 16,
  "model_type": "i2v",
  "num_heads": 12,
  "num_layers": 30,
  "out_dim": 16,
  "patch_size": [
    1,
    2,
    2
  ],
  "qk_norm": true,
  "text_dim": 4096,
  "text_len": 512,
  "window_size": [
    -1,
    -1
  ]
}