Skip to content

Document Question Answering

Eval Recipe for model migration

This Eval Recipe demonstrates how to compare performance of a document question answering prompt with Gemini 1.0 and Gemini 2.0 using Vertex AI Evaluation Service.

  • Use case: answer questions based on information from the given document.

  • The Evaluation Dataset is based on SQuAD2.0. It includes 6 documents stored as plain text files, and a JSONL file that provides ground truth labels: dataset.jsonl. Each record in this file includes 3 attributes:

    • document_path: relative path to the plain text document file
    • question: the question that we want to ask about this particular document
    • reference: expected correct answer or special code ANSWER_NOT_FOUND used to verify that the model does not hallucinate answers when the document does not provide enough information to answer the given question.

  • Prompt Template is a zero-shot prompt located in prompt_template.txt with two prompt variables (document and question) that are automatically populated from our dataset.

  • Python script eval.py configures the evaluation:

    • run_eval: configures the evaluation task, runs it on the 2 models and prints the results.
    • load_dataset: loads the dataset including the contents of all documents.

  • Shell script run.sh installs the required Python libraries and runs eval.py

How to run this Eval Recipe

  1. Configure your Google Cloud Environment and clone this Github repo to your environment. We recommend Cloud Shell or Vertex AI Workbench.
git clone --filter=blob:none --sparse https://github.com/GoogleCloudPlatform/applied-ai-engineering-samples.git && \
cd applied-ai-engineering-samples && \
git sparse-checkout init && \
git sparse-checkout set genai-on-vertex-ai/gemini/model_upgrades && \
git pull origin main
  1. Navigate to the Eval Recipe directory in terminal, set your Google Cloud Project ID and run the shell script run.sh.
cd genai-on-vertex-ai/gemini/model_upgrades/document_qna/vertex_script
export PROJECT_ID="[your-project-id]"
./run.sh
  1. The resulting metrics will be displayed in the script output.
  2. You can use Vertex AI Experiments to view the history of evaluations for each experiment, including the final metrics scores.

How to customize this Eval Recipe:

  1. Edit the Python script eval.py:
    • set the project parameter of vertexai.init to your Google Cloud Project ID.
    • set the parameter baseline_model to the model that is currently used by your application
    • set the parameter candidate_model to the model that you want to compare with your current model
    • configure a unique experiment_name for tracking purposes
  2. Replace the contents of dataset.jsonl with your custom data in the same format.
  3. Replace the contents of prompt_template.txt with your custom prompt template. Make sure that prompt template variables map to the dataset attributes.
  4. Please refer to our documentation if you want to further customize your evaluation. Vertex AI Evaluation Service has a lot of features that are not included in this recipe.