PyTorch code used to train a deep learning model for the diagnosis of Hashimoto's thyroiditis from thyroid ultrasound images.

Code brought from https://github.com/pytorch/vision/tree/master/references/classification.

The example was tested with Python (v3.8.5), PyTorch (v1.7.1+cu101) and TorchVision (v0.8.2) on Ubuntu 16.04.5 and 20.04.2 LTS.

To run the example train.sh for each task, you need to install pytorch and torchvision.
For detail installation procedure, please refer to https://pytorch.org.

The example can be run without GPU. However, GPUs are recommended to use if training with large-scale image data.

1. Installation (5 min)

pip install torch==1.7.1 torchvision==0.8.2 sklearn pandas

2. Git clone (1 min)

git clone https://github.com/lixiangchun/AIplus

3. Run HTNet example (10 min)

cd AIplus/HTNet

cd image-modality
bash train.sh

cd ../multi-modality
bash train.sh

4. Expected output

Model checkpoint file.