config.json

{
  "architectures": [
    "STDiT2"
  ],
  "caption_channels": 4096,
  "class_dropout_prob": 0.1,
  "depth": 28,
  "drop_path": 0.0,
  "enable_flash_attn": true,
  "enable_layernorm_kernel": true,
  "freeze": null,
  "hidden_size": 1152,
  "in_channels": 4,
  "input_size": [
    null,
    null,
    null
  ],
  "input_sq_size": 512,
  "mlp_ratio": 4.0,
  "model_max_length": 200,
  "model_type": "STDiT2",
  "no_temporal_pos_emb": false,
  "num_heads": 16,
  "patch_size": [
    1,
    2,
    2
  ],
  "pred_sigma": true,
  "qk_norm": true,
  "space_scale": 1.0,
  "time_scale": 1.0,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.36.2"
}