DC-DFFN: Densely Connected Deep Feature Fusion Network with Sign Agnostic Learning for Implicit Shape Representation

Reconstruction Preview

Environment

The code is implemented and tested on Ubuntu 20.4 linux environment.

Generation

cd ./code

python evaluate/eval.py --expname shapenet --parallel --exps_folder_name trained_models --timestamp 2022_08_19_16_19_30 --checkpoint 1500 --conf ./confs/shapenet_vae.conf --split ./confs/splits/shapenet/shapenet_sofa_test_files.conf --resolution 100

Training

cd ./code

python training/exp_runner.py --parallel

Acknowledgement

This code is based on SALD (https://github.com/matanatz/SALD), thanks for this wonderful work.