liuzhen03 / AHDRNet

Non-official Pytorch Implementation of Attention-guided Network for Ghost-free High Dynamic Range Imaging (AHDRNet).

AHDRNet

This repository is a PyTorch implementation of CVPR 2019 "Attention-guided Network for Ghost-free High Dynamic Range Imaging"

Results

Todo

Prerequisites

This codebase was developed and tested with pytorch 1.1.1 and Python 3.6.9

• runx for experiments management (you should install runx via source code)
• TensorboardX for training visualization

Training

Data

• which includes 74 samples for training and 15 samples for testing.

Training

To train from terminal, run:

python -m runx.runx experiments/test.yml

Or just run the shell script:

bash run.sh

Run the following commmand for help / more options.

python train -h

To get visualization of the training, you can run tensorboard from the LOGROOT directory using the command:

tensorboard --logdir /data/log/nnhdr --port 6006

Evaluation

Todo