sql-sodabot-v1.0

This encoder-decoder model is a descendent of Salesforce/codet5-small, fine-tuned on a modified version of b-mc2/sql-create-context data. The original CodeT5 was published by Salesfoce Research as an "AI-powered coding assistant to boost the productivity of software developers". The goal of this project is to apply transfer learning in order to appropriate this model for text-to-SQL applications, specifically in the context of generating Socrata SQL (SoQL) queries that can be executed on the Socrata Open Data API (e.g., to analyze NYC Open Data).

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 64
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 25

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu121
• Datasets 2.16.1
• Tokenizers 0.15.0