tencent/hunyuan3d-2mv 🔢❓🖼️✓ → 🖼️

▶️ 7.1K runs 📅 Mar 2025 ⚙️ Cog 0.14.1 🔗 GitHub 📄 Paper ⚖️ License
3d-asset-generation 3d-reconstruction game-asset-generation image-to-3d mesh-generation multi-view multiview

About

Hunyuan3D-2mv is finetuned from Hunyuan3D-2 to support multiview controlled shape generation.

Example Output

Output

Example output

Performance Metrics

14.08s Prediction Time
14.10s Total Time
All Input Parameters
{
  "seed": 1234,
  "steps": 30,
  "file_type": "glb",
  "back_image": "https://replicate.delivery/pbxt/MgQTJZJpLfwXQc6N8rcxSAVGGf2GSGWAFYSqyKQyUVadSfiv/dog2-back.png",
  "left_image": "https://replicate.delivery/pbxt/MgQTKA8In6bsulSoqMd1pp6lhRFmacbXOJuK9swVdknRJSbR/dog2-side.png",
  "num_chunks": 200000,
  "front_image": "https://replicate.delivery/pbxt/MgQTJRO3p8S9ToUYyoMkWLm5edO6y7DZKqP9OZWhYG3zUFtM/dog2-front.png",
  "guidance_scale": 5,
  "randomize_seed": true,
  "target_face_num": 10000,
  "octree_resolution": 256,
  "remove_background": true
}
Input Parameters
seed Type: integerDefault: 1234
Random seed
steps Type: integerDefault: 30Range: 1 - 100
Number of inference steps
file_type Default: glb
Output file type
back_image Type: string
Back view image
left_image Type: string
Left view image
num_chunks Type: integerDefault: 200000Range: 1000 - 5000000
Number of chunks
front_image (required) Type: string
Front view image
right_image Type: string
Right view image
guidance_scale Type: numberDefault: 5
Guidance scale
randomize_seed Type: booleanDefault: true
Randomize seed
target_face_num Type: integerDefault: 10000Range: 100 - 1000000
Target number of faces for mesh simplification
octree_resolution Type: integerDefault: 256Range: 16 - 512
Octree resolution
remove_background Type: booleanDefault: true
Remove image background
Output Schema

Output

Type: stringFormat: uri

Example Execution Logs
Diffusion Sampling::   0%|          | 0/30 [00:00<?, ?it/s]
Diffusion Sampling::   3%|▎         | 1/30 [00:00<00:04,  6.97it/s]
Diffusion Sampling::  10%|█         | 3/30 [00:00<00:02,  9.33it/s]
Diffusion Sampling::  13%|█▎        | 4/30 [00:00<00:03,  8.36it/s]
Diffusion Sampling::  17%|█▋        | 5/30 [00:00<00:03,  7.85it/s]
Diffusion Sampling::  20%|██        | 6/30 [00:00<00:03,  7.55it/s]
Diffusion Sampling::  23%|██▎       | 7/30 [00:00<00:03,  7.35it/s]
Diffusion Sampling::  27%|██▋       | 8/30 [00:01<00:03,  7.23it/s]
Diffusion Sampling::  30%|███       | 9/30 [00:01<00:02,  7.15it/s]
Diffusion Sampling::  33%|███▎      | 10/30 [00:01<00:02,  7.09it/s]
Diffusion Sampling::  37%|███▋      | 11/30 [00:01<00:02,  7.04it/s]
Diffusion Sampling::  40%|████      | 12/30 [00:01<00:02,  7.01it/s]
Diffusion Sampling::  43%|████▎     | 13/30 [00:01<00:02,  6.99it/s]
Diffusion Sampling::  47%|████▋     | 14/30 [00:01<00:02,  6.98it/s]
Diffusion Sampling::  50%|█████     | 15/30 [00:02<00:02,  6.97it/s]
Diffusion Sampling::  53%|█████▎    | 16/30 [00:02<00:02,  6.96it/s]
Diffusion Sampling::  57%|█████▋    | 17/30 [00:02<00:01,  6.95it/s]
Diffusion Sampling::  60%|██████    | 18/30 [00:02<00:01,  6.94it/s]
Diffusion Sampling::  63%|██████▎   | 19/30 [00:02<00:01,  6.94it/s]
Diffusion Sampling::  67%|██████▋   | 20/30 [00:02<00:01,  6.94it/s]
Diffusion Sampling::  70%|███████   | 21/30 [00:02<00:01,  6.93it/s]
Diffusion Sampling::  73%|███████▎  | 22/30 [00:03<00:01,  6.94it/s]
Diffusion Sampling::  77%|███████▋  | 23/30 [00:03<00:01,  6.94it/s]
Diffusion Sampling::  80%|████████  | 24/30 [00:03<00:00,  6.94it/s]
Diffusion Sampling::  83%|████████▎ | 25/30 [00:03<00:00,  6.93it/s]
Diffusion Sampling::  87%|████████▋ | 26/30 [00:03<00:00,  6.93it/s]
Diffusion Sampling::  90%|█████████ | 27/30 [00:03<00:00,  6.93it/s]
Diffusion Sampling::  93%|█████████▎| 28/30 [00:03<00:00,  6.92it/s]
Diffusion Sampling::  97%|█████████▋| 29/30 [00:04<00:00,  6.92it/s]
Diffusion Sampling:: 100%|██████████| 30/30 [00:04<00:00,  6.92it/s]
Diffusion Sampling:: 100%|██████████| 30/30 [00:04<00:00,  7.11it/s]
Volume Decoding:   0%|          | 0/85 [00:00<?, ?it/s]
Volume Decoding:  55%|█████▌    | 47/85 [00:00<00:00, 375.96it/s]
Volume Decoding: 100%|██████████| 85/85 [00:02<00:00, 33.51it/s]
Volume Decoding: 100%|██████████| 85/85 [00:02<00:00, 39.47it/s]
Shape generation took 9.23 seconds
Created new folder: outputs/58ebb0d3-ed6c-4d07-8a2a-f91b81f2b258
Version Details
Version ID
71798fbc3c9f7b7097e3bb85496e5a797d8b8f616b550692e7c3e176a8e9e5db
Version Created
March 18, 2025
Run on Replicate →