The 3d temperature data is gathered from MIT boston and it is used for the purpose of exploring and understanding various deep learning algorithms. Algoritms like MLP, CNN, Conv Autoencoder, and residual neural networks were used in this project. The algorithms were choosen based on the performance of the previous algorithms and the corresponding hyperparameter tuning was done. Among all models, residual networks gave good accuracy and loss. You can find all the models in the form of ipynb files in the repository. best_model.pt can be used if you dont want to run the whole python file. And using this you can load the model settings and the hyperparameters into the python file and get the results. The dataset link is https://drive.google.com/drive/folders/1N-XjQFvzcYUcexQmzcHnckmAFtVzvMsd