Update app.py

Update download_dataset.py

Update README.md

enable rescoring of bvectors

disable rescoring to boost retrieval speed

fix bug

Update app.py

Update search_helpers.py

Update search_helpers.py

change UI, add more info about the process, use binary quantized retrieval

build index on qdrant using binary qunatization

use binary quantization in retrieval + return retrieval/reranking runtime info

Update app.py

Update app.py

Update app.py

Create requirements.txt

Delete requirments.txt

Upload simplewiki-retrieve-rerank.ipynb

Delete simplewiki-retrieve-rerank.ipynb

Update app.py

Create app.py

jupyter notebook for a gradio demo app with full retrieval and reranking functions

helper functions for search relevance: - retrieve top k, rerank hits using cross encoder, extract relevant passage from article, and fetching metadata

code to build the index on qdrant by batches

add requirments.txt file for downloading deps

download the simplewiki dataset and get all articles and passages then deleting the file once the data is ready

initial commit