The method description and the results are provided here.
1.Create a conda environment using requirements.txt using the command provided below.
conda create --name img-classifcation --file requirements.txt
2.requirements.txt already contains Pytorch 1.3 which is supposed to run on a GPU. If you plan to run this code on CPU, install Pytorch using this command. Make sure to be inside the conda environment img-classification while installing it.
```conda install pytorch torchvision cpuonly -c pytorch```
After the requirements are installed, open a terminal to run the model as -
bash scripts/run_<model_name>.sh
Write your model architecture inside models folder. models.softmax.py has been provided for reference.
Scripts inside scripts can be used to train the models. scripts/run_softmax.sh has been provided for reference.