FinQA TAT-QA Financial Finetuning Dataset for Question Answering

FinQA_TAT-QA_financial_finetuning_dataset provides a unified format for question answering over financial documents combining tabular and textual data. The dataset, created by hellotayssir, is built to support training and evaluating models on numerical and discrete reasoning tasks in finance. It draws on the structure and style of established finance-QA benchmarks such as TAT-QA and FinQA.

Use Cases

• Fine-tuning language models for financial question answering based on the described unified context/question/answer format.
• Evaluating model performance on numerical reasoning tasks over financial documents as mentioned in the description.
• Training models to handle hybrid data sources based on the dataset's combination of tabular and textual financial context.

Strengths

• Dataset is built to support training and evaluating models on numerical and discrete reasoning tasks in the finance domain.
• Examples pair a passage of financial context with a question and answer, following a unified, flattened format.
• Dataset draws on the structure and style of established finance-QA benchmarks such as TAT-QA and FinQA.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count, file formats, and license are unknown, which may limit suitability assessment.

Provenance

Source

hellotayssir on Hugging Face.

Collection Method

Built to support training and evaluating models, drawing on established finance-QA benchmarks.

Freshness

Last updated 2026-07-08 13:17:33; freshness should be verified.