ML3

Mini-Project 3

Alexandre Banon; Vincent Delmas; Leila Fabing

Instructions :

You need to add a folder "input" with your *.plk files

Files :

<vgg16.py> to run the vgg16 model and train the weights.

<evaluate.py> to evaluate it.

<lenet5model.py> to run the code that use all of our model from LeNet-5 with modifications.

<svm.py> and <mlp.py> to run the SVM and MLP models.

Python libraries :

vv2, numpy, pandas, sys, matplotlib, torch, skorch, os, sklearn, time, csv, torchvision, pickle, keras, argparse

Links for Dowloading the data :

Modified-Mnist kaggle competitions (february 2019)