ThesisPKinDL

Repository for sharing code from the theses to integrate external resources into neural networks

To run (on python 3)

Install

numpy
pytorch
nltk (for tokenizer), only need the package punkt python -m nltk.downloader 'punkt'
matplotlib

Download

Steps

Run data_tools.py req_embeddings on the glove embeddings and SNLI data to get the subset of embeddings that occur within SNLI
Run Vered's script (https://github.com/vered1986/PythonUtils/blob/master/word_embeddings/convert_text_embeddings_to_binary.py) to create a binary file of the word embeddings
Copy config-template.py to a new file config.py and update the values to correct paths

Name		Name	Last commit message	Last commit date
Latest commit History 1,248 Commits
analyse_repr		analyse_repr
analyse_repr_data		analyse_repr_data
analyses		analyses
clean		clean
models		models
plots		plots
webservices		webservices
.gitignore		.gitignore
README.md		README.md
accuracies_zero_padding		accuracies_zero_padding
analyse.py		analyse.py
config-template.py		config-template.py
create_random_embeddings.py		create_random_embeddings.py
data_tools.py		data_tools.py
embeddingholder.py		embeddingholder.py
eval_twists.py		eval_twists.py
evaluate.py		evaluate.py
find_data.py		find_data.py
find_samples.py		find_samples.py
individual_sentence_models.py		individual_sentence_models.py
log_to_traininfo.py		log_to_traininfo.py
male_female_analyse.py		male_female_analyse.py
model.py		model.py
mydataloader.py		mydataloader.py
pk_analyser.py		pk_analyser.py
pk_experiments.py		pk_experiments.py
plotting.py		plotting.py
rep_act_pair_tools.py		rep_act_pair_tools.py
result_to_conf_matrix.py		result_to_conf_matrix.py
run_model_search.py		run_model_search.py
sent_representation_classify.py		sent_representation_classify.py
swapped_results.txt		swapped_results.txt
test.vocab		test.vocab
train.py		train.py
train_single_model.py		train_single_model.py
visualize_learning.py		visualize_learning.py
word_resource.py		word_resource.py