vufinder/vggt-1b-depth 🖼️✓🔢 → ❓

▶️ 62 runs 📅 Oct 2025 ⚙️ Cog 0.16.8 🔗 GitHub 📄 Paper ⚖️ License

About

Feed-forward neural network that directly infers all key 3D attributes of a scene.

Example Output

Output

{"data":["https://replicate.delivery/czjl/JqZ4lAu6d2atPRleOZ6KPUMjWWh6dZx2m595pLkVQ8NoJqvKA/image_0001.json"],"point_cloud":"https://replicate.delivery/czjl/C3lbkBhIQQ5IB9phggf9Na7rDypVWgmVm3mP0nlC135nJqvKA/point_cloud.glb","depth_images":["https://replicate.delivery/czjl/by90RNWQuQpTONDRIUtFjejnJbpXBNQKponNqpfZRvDQTUfqA/image_0001.png"]}

Performance Metrics

4.00s Prediction Time

172.42s Total Time

All Input Parameters

{
  "images": [
    "https://replicate.delivery/pbxt/NqKIgWflm2NVvlBpLwR3rSlNKogFLy0VGwGI9apml3mtJEUy/0-1.jpg"
  ],
  "return_pcd": true,
  "return_depth": true
}

Input Parameters

inputs (required) Type: array: Input files, accepts JPG, JPEG, PNG, WEBP files for images and MP4, AVI, MOV files for video. The input images and sampled video frames will be padded to a single aspect ratio and resized to a maximum dimension of 518 pixels.
return_pcd Type: booleanDefault: true: Whether to return a point cloud file or not.
return_depth Type: booleanDefault: true: Whether to return depth images or not.
sampling_rate Type: integerDefault: 24: Sampling rate for video input as every n-th frame. Only applies to video inputs. First and last frames of the video will always be included.

Output Schema

Example Execution Logs

Received 1 valid images.
Running inference...
Inference done!
Postprocessing the results...
Building GLB scene...
GLB Scene built!
Dumping data to 1 JSON files...
Dumped 1 JSON files!
Dumping depth data into PNG files...
Dumped 1 PNG files!
Postprocessed the results!

Version Details

Version ID: 0c77fc7a66d501970981d5d9d32c8d0699425c787eb27bfaa9cfa449dfd574ec
Version Created: October 17, 2025

Run on Replicate →