Skip to content

run_table_qa.md

Latest commit

 

History

History
68 lines (52 loc) · 3 KB

run_table_qa.md

File metadata and controls

68 lines (52 loc) · 3 KB

Local Table Question Answering with TAPAS

This script performs Table Question Answering (Table QA) locally using the google/tapas-base-finetuned-wtq model via the Hugging Face transformers library.

You provide a table (as a Pandas DataFrame within the script) and a question about the data in the table, and the model will attempt to find the answer within that table structure.

Features

  • Performs Table QA locally on your machine.
  • Uses the google/tapas-base-finetuned-wtq model (TAPAS architecture).
  • Answers natural language questions based on structured table data.
  • Requires table data defined as a Pandas DataFrame in the script.
  • Leverages the Hugging Face transformers and pandas libraries.
  • Optionally utilizes GPU for faster processing.

Model Used

  • Table QA Model: google/tapas-base-finetuned-wtq (fine-tuned on WikiTableQuestions)

Prerequisites

Before running the script, ensure you have the following installed:

  1. Python: Python 3.8 or later recommended.
  2. System Dependencies: None specific beyond standard build tools.
  3. Python Libraries: You need pandas for table handling and potentially torch-scatter which is often required by TAPAS implementations within transformers. Use pip in a virtual environment: 
    pip install transformers torch pandas torch-scatter
    • transformers: The core Hugging Face library.
    • torch: The deep learning framework backend (PyTorch).
    • pandas: Required for creating and handling the table data structure (DataFrame).
    • torch-scatter: Often a required dependency for TAPAS model computations. Install it explicitly just in case.

Installation

  1. Clone or Download: Get the run_table_qa.py script onto your local machine.
  2. Create Virtual Environment (Recommended): 
    python3 -m venv .venv
source .venv/bin/activate
    (Use .\.venv\Scripts\activate on Windows)
  3. Install Python Libraries: Run the pip command from the Prerequisites section within your activated virtual environment.

Usage

  1. Configure Inputs (Table & Question):

    • Open the run_table_qa.py script in a text editor.
    • Table Data:
      • Locate the section marked # 1. Define the table data....
      • Modify the data dictionary and the resulting pd.DataFrame(data) to represent the table you want to query. Ensure column names are strings.
    • Question:
      • Locate the line: question = "What time is the concert at RAC Arena?"
      • Change the question text to be relevant to the table data you provided.

  2. Run the Script:

    • Open your terminal or command prompt.
    • Make sure your virtual environment is activated.
    • Navigate to the directory containing the script.
    • Execute the script using Python: 
      python run_table_qa.py

Expected Output

The script will print the table data and the question being asked. The final output will be the model's answer derived from analyzing the table: