sssm12 / semantic-similarity-with-bert

Semantic-Similarity-with-BERT

Purpose: This code is to do semantic similarity which means determining the similarity between two sentences. Link to this code: https://github.com/sssm12/Semantic-Similarity-with-BERT

Method: Using SNLI (Stanford Natural Language Inference) corpus to analyze the similarity between sentences with Transformers. For fine-tuning, BERT model is used. This model take inputs of two sentences and outputs the similarity score of those inputs.

Contain three datasets:

• SNLI_Corpus/snli_1.0_train.csv – train_df

• SNLI_Corpus/snli_1.0_dev.csv – valid_df

• SNLI_Corpus/snli_1.0_test.csv – test_df


Process: