Comp 480

Requirements

First, ensure you have the following Python packages installed

• subprocess
• numpy
• matplotlib
• seaborn
• scikit-learn
• math
• random
• pickle
• pandas
• requests
• os
• time
• pillow
• torch
• nearpy

If you are missing any dependencies, please run pip install < missing_dependency_list > to resolve them

Scripts

• run python main.py from the top level directory to generate all results for all parts.

The results will be generated as described below.

LSH PLOTS

• timeByHashFunc.py: Plots the number of hash functions used for LSH vs the average time per query
• errorRates.png: Plots the number of hash functions used for LSH vs the average error rate

CNN METRICS

These are generated to the terminal with every training of the model

• src/cnn/metrics.txt : Contains the training and validation loss per epoch of training, and the time per exhaustive query