How can I obtain image results with 3D bounding boxes?

[robot1lyj]

Thank you for your work!!! I am a novice. I have successfully run the inference code, but there is only the result of one json file. I want to get the picture result with 3D boxes as in the paper. How can I set it up?

[{"answer": "There is one manipulable object part with its 3d bounding box: <box>lid</box>[[012,054,038],[026,049,000],[015,057,040],[036,031,087],[060,025,049],[039,034,089],[057,021,047],[030,053,000]]", "format_answer": [0.3012, 0.54, 0.38, 0.26, 0.49, 0.0, 0.15, 0.57, 0.4, 0.36, 0.31, 0.87, 0.6, 0.25, 0.49, 0.39, 0.34, 0.89, 0.57, 0.21, 0.47, 0.3, 0.53, 0.0], "annotation": null, "question": "A chat between a curious human and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the human's questions.\n\n### Human: Detect all manipulable object parts and provide their 3D bounding boxes.\n### Assistant:", "image": "/home/gpu/A3VLM-main/model/accessory/demo_data/suitcase.png", "fail": true}]