Model Card for ThotaBhanu/t5_sql_askdb

Model Details

Model Description

This model is a T5-based Natural Language to SQL converter, fine-tuned on the WikiSQL dataset. It is designed to convert English natural language queries into SQL queries that can be executed on relational databases.

• Developed by: Bhanu Prasad Thota
• Shared by: Bhanu Prasad Thota
• Model type: T5-based Sequence-to-Sequence Model
• Language(s): English
• License: MIT
• Finetuned from model: t5-large

This model is particularly useful for text-to-SQL applications, allowing users to query databases using plain English instead of writing SQL.

---

Model Sources

• Repository: https://huggingface.co/ThotaBhanu/t5_sql_askdb
• Paper [optional]: N/A
• Demo [optional]: Coming soon

---

Uses

Direct Use

• Convert natural language questions into SQL queries
• Assist in database query automation
• Can be used in chatbots, data analytics tools, and enterprise database search systems

Downstream Use

• Can be fine-tuned further on custom datasets to improve domain-specific SQL generation
• Can be integrated into business intelligence tools for better user interaction

Out-of-Scope Use

• The model does not infer database schema automatically
• May generate incorrect SQL for complex nested queries or multi-table joins
• Not suitable for non-relational (NoSQL) databases

---

Bias, Risks, and Limitations

• The model may not always generate valid SQL for custom database schemas
• Assumes consistent column naming, which may not always be the case in enterprise databases
• Performance depends on how well the input query aligns with the training data format

Recommendations

• Always validate generated SQL before executing on a live database
• Use schema-aware validation methods for production environments
• Consider fine-tuning the model on domain-specific SQL queries

---

How to Get Started with the Model

Use the code below to generate SQL queries from natural language:

from transformers import T5Tokenizer, T5ForConditionalGeneration

# Load model and tokenizer
model_name = "ThotaBhanu/t5_sql_askdb"
tokenizer = T5Tokenizer.from_pretrained(model_name)
model = T5ForConditionalGeneration.from_pretrained(model_name)

# Function to convert query to SQL
def generate_sql(query):
    input_text = f"Convert to SQL: {query}"
    inputs = tokenizer(input_text, return_tensors="pt")
    output = model.generate(**inputs)
    return tokenizer.decode(output[0], skip_special_tokens=True)

# Example usage
query = "Find all employees who joined in 2020"
sql_query = generate_sql(query)

print(f"📝 Query: {query}")
print(f"🛠 Generated SQL: {sql_query}")


## Training Details

### Training Data

Dataset: WikiSQL
Size: 80,654 pairs of natural language questions and SQL queries
Preprocessing: Tokenization using T5Tokenizer, max length 128


### Training Procedure

Training framework: Hugging Face Transformers + PyTorch
Hardware used: NVIDIA V100 GPU
Optimizer: AdamW
Learning rate: 5e-5
Batch size: 8
Epochs: 5

#### Training Hyperparameters

Training precision: Mixed precision (fp16)
Gradient accumulation: Yes (to handle large batch sizes)

#### Speeds, Sizes, Times [optional]

<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->

[More Information Needed]

## Evaluation

<!-- This section describes the evaluation protocols and provides the results. -->

### Testing Data, Factors & Metrics

#### Testing Data

<!-- This should link to a Dataset Card if possible. -->

[More Information Needed]

#### Factors

<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->

[More Information Needed]

#### Metrics

<!-- These are the evaluation metrics being used, ideally with a description of why. -->

[More Information Needed]

### Results

[More Information Needed]

#### Summary



## Model Examination [optional]

<!-- Relevant interpretability work for the model goes here -->

[More Information Needed]

## Environmental Impact

<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->

Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).

- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]

## Technical Specifications [optional]

### Model Architecture and Objective

[More Information Needed]

### Compute Infrastructure

[More Information Needed]

#### Hardware

[More Information Needed]

#### Software

[More Information Needed]

## Citation [optional]

@misc{t5_sql_askdb,
  author = {Bhanu Prasad Thota},
  title = {T5-SQL AskDB Model},
  year = {2025},
  publisher = {Hugging Face},
  howpublished = {\url{https://huggingface.co/ThotaBhanu/t5_sql_askdb}}
}


**BibTeX:**

[More Information Needed]

**APA:**

[More Information Needed]

## Glossary [optional]

<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->

[More Information Needed]

## More Information [optional]

[More Information Needed]

## Model Card Authors [optional]

[More Information Needed]

## Model Card Contact

[More Information Needed]