ndreca/hunyuan3d-2 🔢🖼️❓✓ → ❓

▶️ 27.2K runs 📅 Feb 2025 ⚙️ Cog 0.15.8 🔗 GitHub 📄 Paper ⚖️ License
image-to-3d mesh-generation

About

[Turbo Mode] Scaling Diffusion Models for High Resolution Textured 3D Assets Generation

Example Output

Output

{"mesh":"https://replicate.delivery/xezq/YBBMnm3RV7LRCx8hF6e57V0c1zJZfFjYPLzmK9eNyIwf2dkTB/mesh.glb"}

Performance Metrics

37.72s Prediction Time
139.13s Total Time
All Input Parameters 
{
  "seed": 1234,
  "image": "https://replicate.delivery/pbxt/MVC2B2XKgv4X13qIpW6t2m59EVfY2CqaS9e2CSsWNHPJjQAd/image.png",
  "steps": 50,
  "num_chunks": 200000,
  "max_facenum": 40000,
  "guidance_scale": 5.5,
  "octree_resolution": 512,
  "remove_background": false
}
Input Parameters
seed Type: integerDefault: 1234
Random seed for generation
image (required) Type: string
Input image for generating 3D shape
steps Type: integerDefault: 50Range: 20 - 50
Number of inference steps
num_chunks Type: integerDefault: 200000Range: 10000 - 200000
Number of chunks for mesh generation
max_facenum Type: integerDefault: 40000Range: 10000 - 200000
Maximum number of faces for mesh generation
guidance_scale Type: numberDefault: 5.5Range: 1 - 20
Guidance scale for generation
octree_resolution Default: 512
Octree resolution for mesh generation
remove_background Type: booleanDefault: true
Whether to remove background from input image
Output Schema
mesh Type: stringFormat: uri
Mesh
Example Execution Logs 
Diffusion Sampling::   0%|          | 0/50 [00:00<?, ?it/s]
Diffusion Sampling::   2%|▏         | 1/50 [00:00<00:08,  6.06it/s]
Diffusion Sampling::   6%|▌         | 3/50 [00:00<00:04, 10.66it/s]
Diffusion Sampling::  10%|█         | 5/50 [00:00<00:03, 13.09it/s]
Diffusion Sampling::  14%|█▍        | 7/50 [00:00<00:03, 13.17it/s]
Diffusion Sampling::  18%|█▊        | 9/50 [00:00<00:03, 13.18it/s]
Diffusion Sampling::  22%|██▏       | 11/50 [00:00<00:02, 13.18it/s]
Diffusion Sampling::  26%|██▌       | 13/50 [00:01<00:02, 13.18it/s]
Diffusion Sampling::  30%|███       | 15/50 [00:01<00:02, 13.16it/s]
Diffusion Sampling::  34%|███▍      | 17/50 [00:01<00:02, 13.16it/s]
Diffusion Sampling::  38%|███▊      | 19/50 [00:01<00:02, 13.15it/s]
Diffusion Sampling::  42%|████▏     | 21/50 [00:01<00:02, 13.13it/s]
Diffusion Sampling::  46%|████▌     | 23/50 [00:01<00:02, 13.13it/s]
Diffusion Sampling::  50%|█████     | 25/50 [00:01<00:01, 13.12it/s]
Diffusion Sampling::  54%|█████▍    | 27/50 [00:02<00:01, 13.11it/s]
Diffusion Sampling::  58%|█████▊    | 29/50 [00:02<00:01, 13.11it/s]
Diffusion Sampling::  62%|██████▏   | 31/50 [00:02<00:01, 13.09it/s]
Diffusion Sampling::  66%|██████▌   | 33/50 [00:02<00:01, 13.10it/s]
Diffusion Sampling::  70%|███████   | 35/50 [00:02<00:01, 13.11it/s]
Diffusion Sampling::  74%|███████▍  | 37/50 [00:02<00:00, 13.11it/s]
Diffusion Sampling::  78%|███████▊  | 39/50 [00:03<00:00, 13.10it/s]
Diffusion Sampling::  82%|████████▏ | 41/50 [00:03<00:00, 13.11it/s]
Diffusion Sampling::  86%|████████▌ | 43/50 [00:03<00:00, 13.11it/s]
Diffusion Sampling::  90%|█████████ | 45/50 [00:03<00:00, 13.09it/s]
Diffusion Sampling::  94%|█████████▍| 47/50 [00:03<00:00, 13.10it/s]
Diffusion Sampling::  98%|█████████▊| 49/50 [00:03<00:00, 13.09it/s]
Diffusion Sampling:: 100%|██████████| 50/50 [00:03<00:00, 12.98it/s]
2025-06-21 14:44:29,999 - hy3dgen.shapgen - INFO - FlashVDMVolumeDecoding Resolution: [63, 126, 252, 504]

FlashVDM Volume Decoding:   0%|          | 0/2 [00:00<?, ?it/s]
FlashVDM Volume Decoding: 100%|██████████| 2/2 [00:00<00:00, 29.59it/s]
Decimating mesh... Deleting 993434 faces
Decimation done. Resulting mesh has 40000 faces
Version Details
Version ID
0602bae6db1ce420f2690339bf2feb47e18c0c722a1f02e9db9abd774abaff5d
Version Created
June 21, 2025
Run on Replicate →