ab-bh / Text-similarity-with-SNLI-Network

Semantic similarity analysis between question pairs using GloVe word embeddings on a Stanford Natural Language Inference benchmark model

Text-similarity-with-SNLI

Stanford Natural Language Inference Model

Dataset

Implemented the siamese neural network architecture on Kaggle dataset of Quora Duplicate Question pair on Keras framework.
GloVe 6B token word embeddings.

Results

loss: 0.43
accuracy : 0.79
mean_squared_error: 0.13