Indexnusrefather/Super-Slop-Machina-Macro-Roleplay-Instruct-350M-0812
Indexnusrefather's Super-Slop-Machina-Macro-Roleplay-Instruct-350M-0812 is a 0.35 billion parameter instruction-tuned language model with a 32768 token context length, specifically fine-tuned for enhanced roleplay capabilities. This model is designed to write more consistently, better track plot progression, and handle long-term conversations, making it suitable for local, resource-constrained roleplaying applications.
Loading preview...
Model Overview
Super-Slop-Machina-Macro-Roleplay-Instruct-350M-0812 is a compact 350 million parameter instruction-tuned model developed by Indexnusrefather, featuring a substantial 32768 token context length. It is specifically optimized for local roleplay scenarios, even on devices with limited computational resources.
Key Improvements & Capabilities
This model demonstrates several enhancements over its base version, primarily focused on improving the roleplaying experience:
- Enhanced Roleplay Quality: Significantly better at generating roleplay-centric responses.
- Consistent Writing: Produces more coherent and consistent text outputs.
- Improved Plot Tracking: Shows better ability to maintain awareness of the ongoing plot.
- Long-Term Conversation Handling: Manages extended conversations more effectively.
Training Details
The model was fine-tuned using a full parameter finetuning method with a high learning rate, leveraging a private dataset of chatlogs. This approach led to a drastic change in the model's writing style and expression, specifically tailored for its roleplay objective. Despite these improvements, due to its small size, users should manage expectations regarding its performance compared to much larger models.
Usage Considerations
For quantization, the developer recommends not going lower than Q6_K for optimal performance.