{
  "_class_name": "VidDubModel",
  "format_version": 1,
  "model_type": "viddub",
  "variant": "full_context_2step",
  "causal": false,
  "stage": "dmd",
  "base_model": "Wan2.1-VACE-1.3B",
  "audio_encoder": "openai/whisper-small",
  "fps": 24,
  "sample_rate": 16000,
  "num_frames": 81,
  "latent_chunk_size": 3,
  "num_reference_latents": 1,
  "shift": 5.0,
  "student_steps": 2,
  "inference_audio_cfg": 1.0,
  "training_teacher_cfg": 2.0,
  "sampling": "flow_euler",
  "paper_selected": true,
  "source_run": "FULL2_DMD_R4",
  "trained_component_parameters": 2440958285,
  "parameter_count_scope": "DiT + VACE + audio adapter; external VAE, T5, and Whisper excluded",
  "precision": {
    "dit": "bfloat16",
    "vace": "bfloat16",
    "audio_adapter": "float32"
  },
  "lineage": [
    "r2a EMA@200",
    "FULL2_ODE_R3@400",
    "FULL2_DMD_R4@100"
  ],
  "components": {
    "dit": "dit/diffusion_pytorch_model.safetensors",
    "vace": "vace/diffusion_pytorch_model.safetensors",
    "audio_adapter": "audio_adapter/model.safetensors"
  }
}