mirror of
https://github.com/MervinPraison/PraisonAI.git
synced 2026-11-10 03:31:28 -10:00
44 lines
1.2 KiB
Markdown
44 lines
1.2 KiB
Markdown
# EvaluationLoop Examples
|
|
|
|
Examples demonstrating iterative evaluation and improvement patterns.
|
|
|
|
## Quick Start
|
|
|
|
```bash
|
|
pip install praisonaiagents
|
|
export OPENAI_API_KEY=your-key
|
|
```
|
|
|
|
## Examples
|
|
|
|
| File | Description |
|
|
|------|-------------|
|
|
| `basic_evaluation_loop.py` | Simple loop: Run → Judge → Improve |
|
|
| `multi_step_task_loop.py` | Sequential tasks with context passing |
|
|
| `recipe_optimization_loop.py` | CLI-based recipe optimization |
|
|
|
|
## Pattern
|
|
|
|
```
|
|
┌─────────┐ ┌─────────┐ ┌──────────┐
|
|
│ Agent │ ──► │ Judge │ ──► │ Feedback │
|
|
│ runs │ │ scores │ │ improves │
|
|
└─────────┘ └─────────┘ └──────────┘
|
|
▲ │
|
|
└───────────────────────────────┘
|
|
(repeat until 8.0+)
|
|
```
|
|
|
|
## Run Examples
|
|
|
|
```bash
|
|
# Basic loop
|
|
python basic_evaluation_loop.py
|
|
|
|
# Multi-step with context
|
|
python multi_step_task_loop.py
|
|
|
|
# CLI recipe loop
|
|
python recipe_optimization_loop.py "research AI trends"
|
|
```
|