Skip to content

enesdoruk/DomAtt-UNet

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

7 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Domain Adaptive U-Net: Domain Adaptive Semantic segmentation with PyTorch

Customized implementation of the U-Net with domain adaptive module in PyTorch for Crop And Weed from high definition images.

Quick start

Without Docker

  1. Install CUDA

  2. Install PyTorch 1.13 or later

  3. Install dependencies

pip install -r requirements.txt
  1. Download the data and run training:

For downloading data follow steps on CropAndWeed

python train.py --amp

Description

This model was trained from scratch with 8k images and scored a Dice coefficient of 0.968423 on over 2k test images.

It can be easily used for multiclass segmentation, portrait segmentation, medical segmentation, ...

Usage

Note : Use Python 3.6 or newer

Training

> python train.py -h
usage: train.py [-h] [--epochs E] [--batch-size B] [--learning-rate LR]
                [--load LOAD] [--scale SCALE] [--validation VAL] [--amp]

Train the UNet on images and target masks

optional arguments:
  -h, --help            show this help message and exit
  --epochs E, -e E      Number of epochs
  --batch-size B, -b B  Batch size
  --learning-rate LR, -l LR
                        Learning rate
  --load LOAD, -f LOAD  Load model from a .pth file
  --scale SCALE, -s SCALE
                        Downscaling factor of the images
  --validation VAL, -v VAL
                        Percent of the data that is used as validation (0-100)
  --amp                 Use mixed precision

By default, the scale is 0.5, so if you wish to obtain better results (but use more memory), set it to 1.

Automatic mixed precision is also available with the --amp flag. Mixed precision allows the model to use less memory and to be faster on recent GPUs by using FP16 arithmetic. Enabling AMP is recommended.

Prediction

After training your model and saving it to MODEL.pth, you can easily test the output masks on your images via the CLI.

To predict a single image and save it:

python predict.py -i image.jpg -o output.jpg

To predict a multiple images and show them without saving them:

python predict.py -i image1.jpg image2.jpg --viz --no-save

> python predict.py -h
usage: predict.py [-h] [--model FILE] --input INPUT [INPUT ...] 
                  [--output INPUT [INPUT ...]] [--viz] [--no-save]
                  [--mask-threshold MASK_THRESHOLD] [--scale SCALE]

Predict masks from input images

optional arguments:
  -h, --help            show this help message and exit
  --model FILE, -m FILE
                        Specify the file in which the model is stored
  --input INPUT [INPUT ...], -i INPUT [INPUT ...]
                        Filenames of input images
  --output INPUT [INPUT ...], -o INPUT [INPUT ...]
                        Filenames of output images
  --viz, -v             Visualize the images as they are processed
  --no-save, -n         Do not save the output masks
  --mask-threshold MASK_THRESHOLD, -t MASK_THRESHOLD
                        Minimum probability value to consider a mask pixel white
  --scale SCALE, -s SCALE
                        Scale factor for the input images

You can specify which model file to use with --model MODEL.pth.

Weights & Biases

The training progress can be visualized in real-time using Weights & Biases. Loss curves, validation curves, weights and gradient histograms, as well as predicted masks are logged to the platform.

When launching a training, a link will be printed in the console. Click on it to go to your dashboard. If you have an existing W&B account, you can link it by setting the WANDB_API_KEY environment variable. If not, it will create an anonymous run which is automatically deleted after 7 days.

Data

The Crop and Weed data is available on the Crop And Weed.

The input images and target masks should be in the data/imgs and data/masks folders respectively (note that the imgs and masks folder should not contain any sub-folder or any other files, due to the greedy data-loader).

You can use your own dataset as long as you make sure it is loaded properly in utils/data_loading.py.


Network

Acknowledgments

  • We thank the authors of UNet their open-source codes.