collection of my deep learning models
- tutorialspoint to learn Numpy Pandas ...
- Intro to Deep Learning with PyTorch Udacity course to learn about deep learning
- Kaggle to find datasets and challengs
- Colab to train models in gpu for free :)
-
Digits Recognizer model, from 28x28x1 images (MNIST dataset), the challenge from Kaggle digit-recognizer you can find this model in
./Digits.ipynb -
Titanic3 Dataset analyse and survive prediction using MLP pytorch, I got 84% accuracy :) . you can find the dataset description here and the dataset here the model is in
./Titanic3.ipynb -
Arabic Letters classifier using pytorch (MLP), the data is 32x32x3 images I got 73% accuracy ;_;
the CNN model on the other hand performed much better,the test accuracy is 84.9 (the MLP was trained on 150 epoch)
you can find the MLP model in
./arabic.ipynb,the CNN model in./arabic_CNN.ipynband the data in./Datasets/arabic, the data is structred to be loaded directly to pytorch ImageFolder Dataset -
AI hack tunisia precalificacion is a ML challenge to predict if an african have a bank account, it's also a precalificacion for AI hack tunisia, I got an accuracy of 99.89 %.
you can find the challenge online here
feel free to contribute or to report any error or improvment to any model, i would be happy to add your model here too :)