case.json

{
  "input_mode": "text_prompt",
  "model_source": "/datadrive1/v-zihanyang/diffsynth-studio/train-filtered-1000.safetensors",
  "checkpoint": "/datadrive1/v-zihanyang/diffsynth-studio/train-filtered-1000.safetensors",
  "lora": null,
  "lora_alpha": null,
  "attention_checkpoint": null,
  "attention_mode": "dense",
  "prompt": null,
  "prompts": [
    "A realistic medium close-up of a middle-aged East Asian man in a tidy woodworking studio, warm morning light falling across his face. He looks directly toward the camera and calmly speaks in Mandarin Chinese: ‘准备工作已经完成，我们现在开始吧。’ His mouth shapes and jaw movement are clearly visible and accurately synchronized with the speech. Natural blinking, a small confident nod at the end, stable camera, cinematic detail. Audio: one clear natural male Mandarin voice, close-miked, quiet workshop room tone, no music, no other voices."
  ],
  "original_prompt_count": 1,
  "memory_source": "generated",
  "groundtruth_videos": null,
  "metadata_index": null,
  "cache_path": null,
  "is_train_sample": null,
  "seed": 102,
  "cfg_scale": 3.0,
  "ref_cfg_scale": 1.0,
  "num_inference_steps": 50,
  "height": 480,
  "width": 832,
  "num_frames": 124,
  "shots": [
    {
      "shot_index": 0,
      "path": "shot_00_dense.mp4",
      "prompt": "A realistic medium close-up of a middle-aged East Asian man in a tidy woodworking studio, warm morning light falling across his face. He looks directly toward the camera and calmly speaks in Mandarin Chinese: ‘准备工作已经完成，我们现在开始吧。’ His mouth shapes and jaw movement are clearly visible and accurately synchronized with the speech. Natural blinking, a small confident nod at the end, stable camera, cinematic detail. Audio: one clear natural male Mandarin voice, close-miked, quiet workshop room tone, no music, no other voices.",
      "seed": 102,
      "num_frames": 124,
      "audio_samples": 165600,
      "reference_image_shots": [],
      "reference_audio_shots": [],
      "memory_source": "generated"
    }
  ],
  "generated_path": "multishot_combined_dense.mp4",
  "generated_frames": 124,
  "generated_audio_shape": [
    2,
    165600
  ]
}