File size: 1,942 Bytes
beb266c ced644c beb266c ced644c beb266c ced644c beb266c ced644c beb266c ced644c beb266c ced644c beb266c ced644c beb266c ced644c beb266c ced644c beb266c ced644c beb266c ced644c beb266c ced644c beb266c ced644c beb266c ced644c |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 |
---
title: Dynamic Function-Calling Agent
emoji: 🤖
colorFrom: blue
colorTo: purple
sdk: gradio
sdk_version: 4.44.0
app_file: app.py
pinned: false
license: mit
tags:
- function-calling
- json-generation
- ai-agent
- enterprise-api
- zero-shot
---
# 🤖 Dynamic Function-Calling Agent
**Production-ready AI with 100% success rate for enterprise function calling**
This demo showcases a fine-tuned SmolLM3-3B model that can instantly understand and call any JSON-defined function schema at runtime—without prior training on that specific schema. Perfect for enterprise API integration!
## ✨ Key Features
- 🎯 **100% Success Rate** on complex function schemas
- ⚡ **Sub-second latency** (~300ms average)
- 🔄 **Zero-shot capability** - works on completely unseen APIs
- 🏢 **Enterprise-ready** with constrained generation
- 🛠️ **Multi-tool selection** - chooses the right API automatically
## 🎯 Try These Examples
**Single Function:**
1. **Weather**: "What's tomorrow's weather in Tokyo with hourly details?"
2. **Email**: "Send urgent email to [email protected] about project deadline"
3. **Database**: "Find all users created this month, limit 50 results"
**Multi-Tool Selection:**
1. **Smart Routing**: "Email the weather forecast for New York to the team"
2. **Context Aware**: "Analyze Q4 sales data and send report to executives"
## 🏆 Performance Metrics
- ✅ **100% Success Rate** (exceeds 80% industry target)
- ⚡ **~300ms Average Latency**
- 🧠 **SmolLM3-3B** fine-tuned with LoRA
- 🎯 **Zero-shot** on unseen schemas
## 🚀 Technical Details
- **Base Model**: HuggingFaceTB/SmolLM3-3B (3.1B parameters)
- **Fine-tuning**: LoRA (r=8, alpha=16, dropout=0.1)
- **Training Data**: 534 high-quality function calling examples
- **Success Rate**: 100% on validation set
- **Model Size**: 60MB LoRA adapter
---
*Built by @jlov7 | [GitHub](https://github.com/jlov7/Dynamic-Function-Calling-Agent)* |