It contains two models namely 'model_t1t2.h5' for task1 and task2 and other one 'model_t3.h5' for task3.
The training part is same for task1 and task2 (Training_t1t2.ipynb) which is basically trained on task1 and works good for task2 also.
For task3 we have changed the model to a larger model with a change in some hyperparameters (Training_t3.ipynb).
We have created only one notebook for testing all the three tasks.
Currently Testing.ipynb is ready for task1 and task2.
To test for task3 just change the second argument to 'model_t3.h5' while calling test() in Testing.ipynb.
numpy==1.18.2
gym==0.17.1
Keras==2.3.1
Keras-Applications==1.0.8
Keras-Preprocessing==1.1.0
tensorboard==2.1.1
tensorflow==2.1.0
tensorflow-estimator==2.1.0
matplotlib==3.2.1