| Property | Value |
|---|---|
| Parameter Count | 7.61B |
| Base Model | Qwen/Qwen2.5-Coder-7B-Instruct |
| License | CC-BY-4.0 |
| Paper | Hammer: Robust Function-Calling |
| Tensor Type | BF16 |
What is Hammer2.0-7b?
Hammer2.0-7b is a specialized language model designed for robust function calling capabilities. Built upon the Qwen 2.5 Coder series, it has been fine-tuned using function masking techniques on a dataset of 60,000 APIGen Function Calling samples plus 7,500 custom-generated examples.
Implementation Details
The model utilizes advanced function masking techniques and requires transformers>=4.37.0 for implementation. It's trained on a combination of Salesforce/xlam-function-calling-60k and MadeAgents/xlam-irrelevance-7.5k datasets, optimized for BF16 precision.
- Advanced function masking architecture
- Comprehensive training on API-focused datasets
- Optimized for practical function calling applications
- State-of-the-art performance on BFCL-v3 benchmark
Core Capabilities
- Robust function calling and API integration
- Superior performance on Berkeley Function-Calling Leaderboard
- Efficient parameter utilization at 7.61B scale
- Consistent performance across multiple academic benchmarks
- Flexible integration with Python environments
Frequently Asked Questions
Q: What makes this model unique?
Hammer2.0-7b stands out for its specialized function calling capabilities and consistent performance across benchmarks, outperforming many larger models in its specific domain.
Q: What are the recommended use cases?
The model is ideal for developing personalized, on-device agentic applications, API integration, and scenarios requiring robust function calling capabilities. It's particularly suited for production environments requiring reliable API interaction.