| Property | Value |
|---|---|
| Base Model | Meta-Llama-3.1-8B-Instruct |
| License | Apache 2.0 |
| Training Hardware | Tesla T4 |
| Memory Usage | ~7.922 GB |
What is llama3.1-8b-text2cypher?
llama3.1-8b-text2cypher is a specialized language model fine-tuned from Meta's Llama-3.1-8B specifically for converting natural language queries into Cypher, the query language used by Neo4j graph databases. Developed by Azzedde, this model leverages the Unsloth framework for efficient training and inference, making it particularly valuable for database administrators and developers working with graph databases.
Implementation Details
The model was fine-tuned using the Neo4j Text2Cypher dataset (2024v1), implementing LoRA fine-tuning techniques. The training process utilized specific hyperparameters including a batch size of 2, gradient accumulation steps of 4, and a learning rate of 2e-4, with fp16 precision enabled.
- Efficient implementation using Unsloth framework
- Fine-tuned using LoRA methodology
- Optimized for minimal memory footprint (~7.922 GB)
- Trained on specialized Neo4j query datasets
Core Capabilities
- Natural language to Cypher query conversion
- Database schema interpretation
- Support for complex graph database queries
- Integration with Neo4j database environments
- Automated query generation for knowledge graphs
Frequently Asked Questions
Q: What makes this model unique?
This model specializes in translating natural language into precise Cypher queries, filling a crucial gap in graph database automation. Its fine-tuning on the Neo4j Text2Cypher dataset makes it particularly effective for database-specific tasks.
Q: What are the recommended use cases?
The model excels in database administration, knowledge graph construction, and query automation for structured data retrieval. It's particularly useful for enterprises looking to automate their graph database interactions and enhance semantic search capabilities.