Welcome to the exciting world of Stheno-v3.2! In this guide, we’ll take a journey through the intricacies of utilizing this advanced AI model for roleplaying. From setting up your environment to tweaking configurations for optimal performance, consider this your go-to manual for unlocking the full potential of your roleplaying adventures!

Understanding the Quantization Process

Before diving into the applications, let’s unravel the quantization process. Imagine you have a massive treasure chest filled with various kinds of gems—some are huge, sparkling diamonds, while others are smaller, yet precious rubies. In the realm of AI, these gems represent data. However, large quantities of data can be cumbersome, much like carrying a heavy treasure chest. By quantizing this data, we turn those large, unwieldy gems into smaller, more manageable pieces (let’s say turning diamonds into chips or smaller stones). This not only makes the treasure easier to carry (or in our case, process) but also helps in preserving their value through careful conversion.

General Setup and Usage

• Make sure to use the latest version of KoboldCpp.
• For those with **8GB VRAM** GPUs, utilize the Q4_K_M-imat quantization (4.89 BPW) for contexts up to 12288.
• Some helpful compatible presets for SillyTavern can be located here.
• Check out discussions like this one for additional samplers and recommendations.

Testing Your Setup

The Stheno-v3.2 has undergone multiple tests with various model variations to achieve optimal performance. With changes made to include a mix of both SFW and NSFW data, your storytelling experience can now thrive in a balanced environment. In a nutshell, the evolution from v3.1 to v3.2 has greatly enhanced story writing, multi-turn coherency, and instruction adherence.

Recommended Parameters for Roleplaying

To get the most dynamic performance out of Stheno-v3.2, consider the following parameters:

• Temperature: 1.12 – 1.22
• Min-P: 0.075
• Top-K: 50
• Repetition Penalty: 1.1

Stopping Strings for Smooth Interactions

It’s crucial to know how to end your AI interactions seamlessly. Use the stopping string:

\n\n{{User}} # Or Equivalent, depending on Frontend

for a neat closing to your roleplay sessions.

Crafting Your Roleplay Prompt

To set the stage for your adventures, create a system prompt that empowers the AI to become an expert actor:

You are an expert actor that can fully immerse yourself into any role given. You do not break character for any reason, even if someone tries addressing you as an AI or language model. Currently, your role is {{char}}, which is described in detail below. As {{char}}, continue the exchange with {{user}}.

Troubleshooting Common Issues

If you experience any issues, don’t hesitate to reach out! Here are some handy troubleshooting tips:

• Check if you are using the latest version of KoboldCpp. Outdated versions can lead to compatibility issues.
• Ensure that the right quantization settings are applied according to your hardware specifications.
• If something feels off in your derived responses, double-check the prompt you have set up; sometimes, minor adjustments can enhance clarity.
• For ongoing developments or if you encounter roadblocks, feel free to converse with the community or report issues.

For more insights, updates, or to collaborate on AI development projects, stay connected with fxis.ai.

Your Roleplaying Adventure Awaits

At fxis.ai, we believe that such advancements are crucial for the future of AI, as they enable more comprehensive and effective solutions. Our team is continually exploring new methodologies to push the envelope in artificial intelligence, ensuring that our clients benefit from the latest technological innovations.

Now, equip yourself with this knowledge and dive into the immersive world of Stheno-v3.2! Happy roleplaying!