Demo

VARAG Plaground¶

Explore VARAG with our interactive playground! It lets you seamlessly compare various RAG (Retrieval-Augmented Generation) solutions, from data ingestion to retrieval.

This makes it easy to test and experiment with different approaches in real-time.

Recommended: Run this Notebook with a T4 GPU

In [ ]:

!git clone https://github.com/adithya-s-k/VARAG
%cd VARAG
%pwd

!git clone https://github.com/adithya-s-k/VARAG %cd VARAG %pwd

In [ ]:

!apt-get update && apt-get install -y && apt-get install -y poppler-utils

!apt-get update && apt-get install -y && apt-get install -y poppler-utils

In [ ]:

%pip install -e .

## We will be using Docling for OCR
%pip install docling

%pip install -e . ## We will be using Docling for OCR %pip install docling

In [ ]:

!pip install --upgrade gradio

!pip install --upgrade gradio

Add your OpenAI API Key

To add your OpenAI API key and Gemini API Key to Colab secrets:

1. Click on the "🔑 Secrets" tab in the left sidebar of your Colab notebook.
2. Click the "+" button to add a new secret.
3. In the "Name" field, enter OPENAI_API_KEY.
4. In the "Value" field, paste your OpenAI API key.
5. Make sure the "Notebook access" toggle is enabled for this secret.
6. Repeat the same process for GEMINI_API_KEY.
7. Click "Done".

Now your keys are securely stored and can be accessed in your notebook using userdata.get('OPENAI_API_KEY').

In [ ]:

import os
from google.colab import userdata

# Store your OpenAI API key in Colab secrets under the name 'OPENAI_API_KEY'
#os.environ['OPENAI_API_KEY'] = userdata.get('OPENAI_API_KEY') #USES OPENAI
os.environ['GEMINI_API_KEY'] = userdata.get('GEMINI_API_KEY') #USES LITELLM

import os from google.colab import userdata # Store your OpenAI API key in Colab secrets under the name 'OPENAI_API_KEY' #os.environ['OPENAI_API_KEY'] = userdata.get('OPENAI_API_KEY') #USES OPENAI os.environ['GEMINI_API_KEY'] = userdata.get('GEMINI_API_KEY') #USES LITELLM

Run Gradio¶

In [ ]:

#This is the code to run the VARAG Demo without any Interpretability maps
!python demo.py --share

#This is the code to run the VARAG Demo without any Interpretability maps !python demo.py --share

For exploring colpali interpretation check out codes in examples/inference_colpali

In [ ]:

#Run this to load the Interpretability maps feature for colpali
!python examples/inference_colpali/demo_with_heatmaps.py --share

#Run this to load the Interpretability maps feature for colpali !python examples/inference_colpali/demo_with_heatmaps.py --share