# elsciRL GUI Application Guide
<div align="center">
<iframe width="700" height="400"
src="https://www.youtube.com/embed/JbPtl7Sk49Y">
</iframe>
</div>
## Overview
The elsciRL GUI is a web-based application that allows you to run reinforcement learning experiments with various agents and adapters. The interface is organized into five main tabs, each serving a specific purpose in the experiment workflow.
## Step-by-step Guide
### 1. Start with Application Selection
- Choose an application that matches your research goals
- Review preview images to understand the problem structure
- Select appropriate configuration options
### 2. Configure Your Experiment
- Import preset configurations when available
- Set reasonable training/testing parameters
- Select multiple agents for comparison
- Configure agent parameters based on problem characteristics
### 3. Provide Clear Instructions
- Use the LLM Instruction Planner for validation
- Write specific, actionable instructions
- Validate agent understanding before training
- Iterate on instructions if needed
### 4. Train Your Agent and Track Progress
- Watch real-time updates during training
- Note any warnings or errors
- Allow sufficient time for completion
### 5. Analyze Results Thoroughly
- Review all generated figures and metrics
- Compare performance across agents
- Consider variance in results
- Save configurations for reproducibility