camenduru/dust3r 🔢🖼️✓❓ → 🖼️
About
DUSt3R: Geometric 3D Vision Made Easy
Example Output
Output
Performance Metrics
4.07s
Prediction Time
390.10s
Total Time
All Input Parameters
{
"niter": 300,
"refid": 0,
"image1": "https://replicate.delivery/pbxt/KVCgnEVoTYocNCeWjCeEjc1RILo8u4d3jqPX9Srak3QiX0rB/frame01.jpg",
"image2": "https://replicate.delivery/pbxt/KVCgnxgya22Ksw8WwG2gYLZIu07Ch6eQkzwoQQDeMbH2FXf4/frame02.jpg",
"winsize": 1,
"cam_size": 0.05,
"mask_sky": false,
"schedule": "linear",
"image_size": 512,
"clean_depth": true,
"min_conf_thr": 3,
"as_pointcloud": false,
"scenegraph_type": "complete",
"transparent_cams": false
}
Input Parameters
- niter
- refid
- image1 (required)
- Input Image 1
- image2 (required)
- Input Image 2
- winsize
- cam_size
- mask_sky
- schedule
- image_size
- clean_depth
- min_conf_thr
- as_pointcloud
- scenegraph_type
- transparent_cams
Output Schema
Output
Example Execution Logs
>> Loading a list of 2 images - adding /tmp/tmp5yedrjsgframe01.jpg with resolution 1042x1854 --> 288x512 - adding /tmp/tmpjgnuflwiframe02.jpg with resolution 1042x1854 --> 288x512 (Found 2 images) >> Inference with model on 2 image pairs 0%| | 0/2 [00:00<?, ?it/s] 50%|█████ | 1/2 [00:01<00:01, 1.05s/it] 100%|██████████| 2/2 [00:01<00:00, 1.50it/s] 100%|██████████| 2/2 [00:01<00:00, 1.38it/s] - conf=86.5 for edge 0-1 - conf=1.1e+02 for edge 1-0 (exporting 3D scene to /content/scene.glb )
Version Details
- Version ID
d0e12b55655b28ed3349cca6d7c567db6237382c59d2e3d5dbe4d324c6df20c7- Version Created
- March 3, 2024