https://github.com/abhijitramesh/semantic-similarity
This repository will contain state of the art implementation of semantic-similarity
https://github.com/abhijitramesh/semantic-similarity
bert huggingface nlp sematic-similarity snli t5
Last synced: 4 months ago
JSON representation
This repository will contain state of the art implementation of semantic-similarity
- Host: GitHub
- URL: https://github.com/abhijitramesh/semantic-similarity
- Owner: abhijitramesh
- Created: 2021-01-12T07:09:47.000Z (over 4 years ago)
- Default Branch: master
- Last Pushed: 2021-01-13T10:30:44.000Z (over 4 years ago)
- Last Synced: 2024-12-29T23:42:27.242Z (6 months ago)
- Topics: bert, huggingface, nlp, sematic-similarity, snli, t5
- Language: Jupyter Notebook
- Homepage:
- Size: 37.1 KB
- Stars: 2
- Watchers: 2
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: readme.md
Awesome Lists containing this project
README
# Semantic Similarity
This repository would contain implementation of semantic similarity with different state of the art transformers from [huggingface](https://huggingface.co/)
## Introduction
Semantic similarity is used to measure the distance or similarity between two pair of words, phrases, sentences or documents. This can be done in two ways knowledge-based or corpus-based. Here we would be doing a corpus-bases approach. This implementation is inspired from semantic similarity implementation with BERT example from [keras.io](https://keras.io/examples/nlp/semantic_similarity_with_bert/). The idea here is to change the models and see if we could get a better performance with other transformer architecture like t5.
### References
[BERT](https://arxiv.org/pdf/1810.04805.pdf)
[T5](https://arxiv.org/abs/1910.10683)
[SNLI](https://nlp.stanford.edu/projects/snli/)