README.md

U-Net Transformer

Unofficial Pytorch implementation of following papers for predicting Protein Subcellular Localization from quantitative label-free imaging with phase and polarization:

• U-Net Transformer
• microDL

Quick Start

1. Create Environment

conda create -n <environment_name> python=3.8
conda activate <enviroment_name>
pip install -r requirements.txt

2. Download dataset as mentioned in Data section
3. Run inference on pretrained weights

python inference.py --config 

Training

python train_test.py --config

For Tensorboard: tensorboard --logdir logs/

Data

QLIPP can be downloaded from repo.

• The directory structure of the whole project is as follows:

.
├── Network
│   ├──datasets
│   │       └── dataset_*.py
│   ├──train.py
│   ├──test.py
│   └──...
├── model
│   └── TU_Synapse128
│       └── res_True_head_4_ch_512_nuclei
│           ├── UTransform-129.pth
│           └── *.pth
└── data
    └──Synapse
        ├── train
        │   ├── im_c001_z011_t000_p005_r0-256_c0-256_sl0-3.npy
        │   └── *.npy
        └── train_label
            ├── im_c000_z011_t000_p005_r0-256_c0-256_sl0-3.npy
            └── *.npy

---

References

• TransUNet
• TransFuse
• Road Extraction Using PyTorch
• Deep Residual-Unet
• Squeeze-and-Excitation Networks
• ResUNet++
• Unet
• Brain tumor segmentation# UTransformer