Home

Awesome

SketchSampler

<p align="center"> <a href="https://pytorch.org/get-started/locally/"><img alt="PyTorch" src="https://img.shields.io/badge/PyTorch-orange?logo=pytorch"></a> <a href="https://pytorchlightning.ai/"><img alt="Lightning" src="https://img.shields.io/badge/-Lightning-blueviolet"></a> <a href="https://hydra.cc/"><img alt="Conf: hydra" src="https://img.shields.io/badge/conf-hydra-blue"></a> <a href="https://wandb.ai/site"><img alt="Logging: wandb" src="https://img.shields.io/badge/logging-wandb-yellow"></a> <a href="https://github.com/lucmos/nn-template"><img alt="Template: nn-template" src="https://shields.io/badge/-nn--template-emerald?style=flat&logo=github&labelColor=gray"></a> </p>

Official PyTorch implementation of paper SketchSampler: Sketch-based 3D Reconstruction via View-dependent Depth Sampling, accepted by ECCV 2022.

Environments

Training & Evaluation

tar zxvf sketchsampler_dataset.tar.gz
python ./src/trainval.py

Configuration

We use hydra to manage the configurations:

.
├── conf
   ├── sketchsampler.yaml    # experiment configuration
   ├── data                  # dataset configuration(num_workers, batch_size, etc.)
   ├── hydra                 # job configuration
   ├── logging               # logging configuration(logging frequency, wanbd configuration, etc.)
   ├── model                 # model configuration
   ├── optim                 # optimizer configuration(optimizer type, learning rate, etc.)
   └── train                 # training configuration(gpus, loss weights, epochs, etc.)

Please feel free to modify any configuration to achieve better results!

Logging

We adopt Weights and Biases as the logging framework. If you want to analyze the results, please install wandb following official document and replace USER in ./config/logging/default.yaml by your own username. export WANDB_MODE=dryrun in .env can be also modified and more options can be found here.

Contact

If you have any questions about the implementation or the paper, please feel free to open an issue or contact me at gaochenjian@buaa.edu.cn.