The main object in this repository is the qa_engine.ipynb
notebook.
It contains a presentation showing how to run a large language model (LLM) on a laptop using LangChain and then index Web documents, to query their content with natural language.
There are different ways to view the presentation:
- 🐍 create environment with
poetry install
, then load the notebook withjupyter notebook qa_engine.ipynb
and pressAlt+R
to enter the presentation mode, ▶️ open the notebookqa_engine.ipynb
in GitLab/GitHub,- 📥 download one of the
exported_*.html
files (if you choose the "slides" file, you also to download theimages/
directory).
Jupyter Notebook v6 is used, because the RISE extension does not work well with Jupyter Lab or v7. This may change when jupyterlab-contrib/rise becomes more mature.
jupyter contrib nbextension install --user
jupyter nbextension enable execute_time/ExecuteTime
jupyter nbextension enable export_embedded/main
jupyter nbconvert qa_engine.ipynb --to html_embed --output ./exported_qa_engine.one_page.html
# export_embedded does not add the option to export slides with embedded images
jupyter nbconvert qa_engine.ipynb --to slides --output ./exported_qa_engine