Skip to content

Deep Learning Lab: Language Modeling with an LSTM, Mathematical problem solving using a Transformer.

Notifications You must be signed in to change notification settings

felixboelter/Deep-Learning-Lab

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

20 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Deep Learning Lab

Installation

Conda:

$ git clone https://github.com/felixboelter/Deep-Learning-Lab
$ conda config --append channels pytorch
$ conda create --name <env_name> --file requirements.txt

Pip:

$ git clone https://github.com/felixboelter/Deep-Learning-Lab
$ pip install torchtext==0.10.0 torchvision==0.10.0 torch==1.9.0 tqdm jupyter numpy matplotlib

Using a convolutional neural network (CNN) to classify the images in the CIFAR-10 dataset. The model includes Convolution layers accompanied with maxpooling layers, finally the model uses fully connected layers to obtain a lower dimensional output of 10 features corresponding to each class in the dataset. PyTorch and Torchvision is used to train this neural network.

Implemented a character-level language model using a recurrent neural network (RNN), specifically the long short-term memory (LSTM) model. The model is trained on the Aesop's Fables from Project Gutenberg. In addition, it is also trained on Donald Trump's rally speeches, to validate the results. The implementation, training procedure, training losses and hyperparameters can be seen in the Report.

Implemented the Transformer sequence-to-sequence model to solve mathematical problems based on the DeepMind Mathematics Dataset which includes three difficulties. A further review of the implementation, hyperparameter tuning, the training and validation losses and their respective accuracies can be seen in the Report.

About

Deep Learning Lab: Language Modeling with an LSTM, Mathematical problem solving using a Transformer.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published