Skip to content

FAQ

Jump to bottom
Jazmia Henry edited this page Sep 7, 2024 · 1 revision

IsoZero Troubleshooting and FAQ

This page addresses common issues you might encounter while using IsoZero and provides answers to frequently asked questions. If you don't find your issue or question addressed here, please check our GitHub issues or reach out to our community support channels.

Table of Contents

Troubleshooting

Installation Issues

  1. Problem: ModuleNotFoundError: No module named 'isozero' Solution: Ensure you've installed IsoZero correctly. Try reinstalling:

    pip uninstall isozero
pip install isozero

  2. Problem: Version conflicts with dependencies Solution: Create a new virtual environment and install IsoZero there:

    python -m venv isozero_env
source isozero_env/bin/activate  # On Windows, use `isozero_env\Scripts\activate`
pip install isozero

API Key Issues

  1. Problem: AuthenticationError: Invalid API key Solution: Double-check that you've set your API key correctly: 
    import os
os.environ['OPENAI_API_KEY'] = 'your-actual-api-key'
# Or for Anthropic:
os.environ['ANTHROPIC_API_KEY'] = 'your-actual-api-key'

Performance Issues

  1. Problem: Reasoning steps are taking too long Solution: Try reducing the number of steps or using a faster model. For example:

    simulation = ReasonSimulation("Your task", max_steps=3)  # Reduce max_steps
# Or use a faster model:
agent = OpenAIAgent(api_key="your-key", model="gpt-3.5-turbo")

  2. Problem: Out of memory errors with large documents Solution: Try processing documents in smaller chunks:

    loader = DocumentLoader(chunk_size=1000)  # Adjust chunk size as needed

Output Quality Issues

  1. Problem: Inconsistent or irrelevant outputs Solution: Try increasing the number of steps or using a more capable model: 
    simulation = ReasonSimulation("Your task", max_steps=5)  # Increase max_steps
# Or use a more capable model:
agent = OpenAIAgent(api_key="your-key", model="gpt-4-turbo")

Frequently Asked Questions

  1. Q: What's the difference between ReasonSimulation and MathSimulation? A: ReasonSimulation is for general reasoning tasks and can handle a wide variety of problems. MathSimulation is optimized for mathematical problem-solving and includes additional steps for equation formulation and calculation.

  2. Q: Can I use IsoZero with my own custom LLM? A: Yes, you can create a custom Agent class that interfaces with your LLM. Inherit from the base Agent class and implement the required methods.

  3. Q: How many steps should I use for my simulation? A: IsoZero simulations have a default number of steps based on the complexity of the task. If you're unsure, you can rely on these defaults as they are designed to handle most common scenarios effectively. However, you can always customize the number of steps:

    • For simple tasks, the default is usually 1-2 steps.
    • For more complex tasks, it might be 3-5 steps.
    • If you find the results unsatisfactory, you can increase the number of steps.
    • For very complex tasks, you might want to set it higher, up to 6+ steps.
    • Keep in mind that most tasks take 3 steps.

    Here's how you can set a custom number of steps:

    simulation = ReasonSimulation("Your task", max_steps=3)  # Set to 3 steps

    Remember, more steps can lead to more thorough reasoning but also increase computation time and cost. It's about finding the right balance for your specific use case.

  4. Q: Can IsoZero handle multiple languages? A: Yes, IsoZero's capabilities depend on the underlying LLM. Most supported models can handle multiple languages, but performance may vary.

  5. Q: Is my data safe when using IsoZero? A: IsoZero itself doesn't store your data. However, when using external LLM APIs, refer to their respective privacy policies. For sensitive data, consider using a local model or a provider with appropriate data handling practices.

  6. Q: Can I use IsoZero offline? A: Yes, if you use it with a local Hugging Face model. However, for Claude or GPT models, an internet connection is required to access their APIs.

  7. Q: How can I contribute to IsoZero? A: We welcome contributions! Check our Contributing Guide for details on submitting pull requests, reporting bugs, or suggesting enhancements.

  8. Q: Is IsoZero free to use? A: IsoZero itself is open-source and free to use. However, you may incur costs from the LLM providers (like OpenAI or Anthropic) based on your usage.

  9. Q: How often is IsoZero updated? A: We strive to update IsoZero regularly with bug fixes, improvements, and new features. Check our GitHub repository for the latest releases and changelog.

  10. Q: Can IsoZero generate images or process audio? A: Currently, IsoZero focuses on text-based tasks. It doesn't have built-in capabilities for image generation or audio processing, although you could potentially use it to generate prompts for image generation models.

If you're still experiencing issues or have a question not answered here, please:

  1. Check our GitHub Issues to see if it's a known problem or has been discussed before.
  2. If not, feel free to open a new issue with a detailed description of your problem or question.
  3. For general discussions, join our community forums or Discord channel (links available on our GitHub page).

We're here to help you make the most of IsoZero!

Clone this wiki locally