search

hzxie / Pix2Vox

The official implementation of "Pix2Vox: Context-aware 3D Reconstruction from Single and Multi-view Images". (Xie et al., ICCV 2019)
https://haozhexie.com/project/pix2vox
MIT License
473 stars 115 forks source link

pix3d models are .obj not .binvox #114

Closed shayansep closed 1 month ago

shayansep commented 1 month ago

hello,

is the code writtent in a way, to train on all the pix3d, pascal3d and shapenet all at once ?

you know that the pix3d 3d-models are in .obj format, but inside the config.py, it asks for binovax formats: __C.DATASETS.PIX3D.VOXEL_PATH = '/path/to/Datasets/Pix3D/model/%s/%s/%s.binvox'

do i have to turn them into binovax ? ( which will be veryyy time consuming ) or its fine like that? ( to keep them as .obj ) and do i have to change the address to : __C.DATASETS.PIX3D.VOXEL_PATH = '/path/to/Datasets/Pix3D/model/%s/%s/%s.obj' ??

and also, you didnt put the link to the specific pascal3d dataset that we have to use, because config.py has also lines for pascal3d paths. and again im asking if the train.py code written in a way to train on all those dataset all at once ?

please guide me thank you for your attention

hzxie commented 1 month ago

Please refer to the answers in the following issues.

  1. 25

  2. 4