You can use conda to create environment
conda create --name imageclassification python=3.6
And you need to activate it
conda activate imageclassification
After that you can install package below with pip or conda install
- Python (3.6)
- Flask
- Pytorch, torchvision (https://pytorch.org/)
- imageio
- You can download from my drive (https://bit.ly/34mN2Cx) and put it all to /server/checkpoints/ folder.
- You can download from here https://www.kaggle.com/tongpython/cat-and-dog
Download npm and nodejs (https://nodejs.org/)
In the root directory of the project...
- Install node modules
yarn install
ornpm install
. - Start development server
yarn start
ornpm start
.
If you want to compare the accuracy between models, you can run file model.py
usage: model.py [-h] [--dataset DATASET] [--checkpoints CHECKPOINTS]
optional arguments:
-h, --help show this help message and exit
--dataset DATASET Dataset test path to evaluation
--checkpoints CHECKPOINTS
Path folder contains checkpoint