bhanu77prakash / EDA-project

EDA-project

Requirements

1. python 3
2. PyTorch

Please note the versions to be installed

torch==0.1.11._5
torchvision==0.1.9

For training

1. Download the data into train_module/data
2. Fowllow the train_module/preprocess/Readme.md
3. Follow the train_module/Readme.md to run train_model.py file.

Pretrained models

The models are present in a drive link mentioned below

https://drive.google.com/drive/folders/1xH3zcEEbIePlnW8G029Lqe4nXeWp5AOa?usp=sharing

Download them and paste them in test_module/models

For testing

Using pretrained models

1. Download the models from the drive link given above.
2. Install the requirements by running pip install -r requirements.txt
3. Inside the test_module run the line python3 FE.py for frontend based model
4. Run the following command for graphical based model

python run_model.py --image img/CLEVR_val_000013.png --question "Does the small sphere have the same color as the cube left of the gray cube?"

For frontend model please use sudo apt-get install python-qt4 to install pyqt-4 and sudo apt install python3-tk to install tkinter

Training from scratch

1. Follow the training section to train the model
2. Then copy the trained models instead of downloading in the above testing step.
3. Follow the rest of the steps mentioned in the testing section

References

1. Stackoverflow
2. Clevr-iep[https://github.com/facebookresearch/clevr-iep]
3. Clevr-iep[https://arxiv.org/abs/1705.03633]