dcharatan / pixelsplat

[CVPR 2024 Oral, Best Paper Runner-Up] Code for "pixelSplat: 3D Gaussian Splats from Image Pairs for Scalable Generalizable 3D Reconstruction" by David Charatan, Sizhe Lester Li, Andrea Tagliasacchi, and Vincent Sitzmann
http://davidcharatan.com/pixelsplat/
MIT License
864 stars 60 forks source link

How to use it? #52

Closed YZWYD closed 5 months ago

YZWYD commented 7 months ago

Hello author, could you please advise me on how to utilize the pre-trained weights you shared to process a pair of images and generate the corresponding 3D model?

YZWYD commented 7 months ago

So can you use the Pre-trained Models now? I cannot see your reply on github.

Z&Y @.***

 

------------------ 原始邮件 ------------------ 发件人: "dcharatan/pixelsplat" @.>; 发送时间: 2024年3月6日(星期三) 凌晨3:12 @.>; @.**@.>; 主题: Re: [dcharatan/pixelsplat] How to use it? (Issue #52)

same question

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>

dcharatan commented 5 months ago

See this part of the README for instructions on how to run evaluation/inference. If you want 3D models in .ply format, you can check out src/paper/generate_point_cloud_figure.py. It uses the export_ply function from src/model/ply_export.py to export the output (3D Gaussians) of the encoder.