Skip to content
Explore different losses used in image inpainting based on deep learning
Python
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
.idea
README.md
benchmark.json
common.py
dataset.py
get_metadata.py
loss_function.py
model.py
model_deep.py
predict.py
train_L1.py
train_L2.py
train_content.py
train_content_mstyle_mtv.py
train_content_tv.py
train_content_tv_mstyle.py
train_style.py
train_style_content.py
train_style_content_tv.py
utils.py

README.md

Note

  • Benchmark: contains 100 images for testing
  • predict.py: run to test model performance on benchmark

Run

To test model on benchmark, one can use command:

python3 predict.py --model_dir=<your model path> --save_dir=<directory to save results>

For example,

CUDA_VISiBLE_DEVICES=0 python3 predict.py --model_dir=models/L1+content+tv.pth.tar --save_dir=results

The default arguement is --model_dir=models/L1+content+tv.pth.tar --save_dir=results

So you can simply run CUDA_VISiBLE_DEVICES=0 python3 predict.py

You can’t perform that action at this time.