tiny-imagenet-classification

Installation

1. Make sure thaat you are in clean enviroment (conda/venv/etc)
2. pip install -r requirements.txt
3. Make sure that you've got torch which is compatible to your CUDA, if you are going to use GPU.

Configuration

1. Go to ./configs
2. Use default_config.yaml and set your own. Models from torchvision you can use: resnet-family, vgg-family, squeezenet-family, densenet-family, mnasnet-family.
3. If you are going to use GPU-training, set their indices as an yaml-array: [0, 1].

Training

python train.py --config=<your_config> --experiments_dir=../<yout_experiments_folder> --experiment_name=<your_experiment_name>

Evaluation

Example is in notebooks/evaluation.ipynb

Pretrained models

You can download experiments with resnet18/50, vgg19, mnasnet1.0 here: https://drive.google.com/file/d/1CAseZTl54txzH6TWL1rUf6NAAiKHUhpB/view?usp=sharing. You can untar them and use as shown in notebooks/evaluation.ipynb.