Hands-on workshop for building a semantic search engine.
During the workshop, use this custom JupyterHub, which has all dependencies preinstalled.
The repo is located at npatta01/search-engine-workshop
To use this repository outside a workshop, use Binder:
Data fetching
These notebooks download the Unsplash dataset and save it in Hugging Face dataset format.
Non-deep-learning retrieval
BM25 retrieval with Elasticsearch: notebook
Deep-learning retrieval (text)
Text-based deep-learning retrieval: notebook
Deep-learning retrieval (image)
CLIP retrieval: notebook
ANN
Explore approximate nearest-neighbor indexes for faster deep-learning retrieval: notebook
For help or feedback, please reach out to:
This workshop uses the Unsplash Lite Dataset 1.2.0.
The hands-on portion of the workshop was made possible by the JupyterHub Helm Chart.
v1.1
- Set up for PyData NYC
- Replaced Stack Overflow data with Unsplash data
v1.0
- Set up for ODSC
- Used Stack Overflow data