tdoehmen/t5-schemapile-fk

SchemaPile Foreign Key Detection Model (T5-base)

Model Description

In this repository we are introducing t5-schemapile-fk. It's a language model, based on google-t5/t5-base fine-tuned for predicting foreign key relationships in relational database schemas.

Training Data

Forein key pairs extracted from SchemaPile-Perm, a large collection of relational database schemas.

Evaluation Data

We evaluate the foreign key detection accuracy of starcoder-schemapile-fk and t5-schemapile-fk on schemas from Spider, BIRD-SQL, and CTU PRLR.

Training Procedure

The model was trained using the following hyperparamters:

• batch_size = 16
• learning_rate=4e-5,
• weight_decay=0.01,
• num_train_epochs=1

See Training Code.

How to Use

We recommend using the following prompt template:

Example Prompt:

You are given the following SQL database tables: 
staff(staff_id, staff_address_id, nickname, first_name, middle_name, last_name, date_of_birth, date_joined_staff, date_left_staff)
addresses(address_id, line_1_number_building, city, zip_postcode, state_province_county, country)
Output a json string with the following schema {table, column, referencedTable, referencedColumn} that contains the foreign key relationship between the two tables.

Example Output:

{'table': 'staff',
 'column': 'staff_address_id',
 'referencedTable': 'addresses',
 'referencedColumn': 'address_id'}

To run the model locally, we recommend using our end-to-end Example Notebook.