Captain-Eris_Violet-GRPO-v0.420 Open-source Language Model - Supports Role-playing and Dialogue Generation Scenarios

Captain Eris Violet GRPO V0.420

Developed by Nitral-AI

Captain-Eris_Violet is an advanced language model developed through multi-stage supervised fine-tuning, QLoRA adapters, and GRPO-optimized RLHF, suitable for role-playing and dialogue generation.

Large Language Model

Transformers

EnglishOpen Source License:Other #Multi-stage RLHF optimization #Character dialogue generation #SillyTavern integration

Downloads 1,355

Release Time : 2/17/2025

Model Overview

This model combines supervised fine-tuning, QLoRA adapters, and GRPO-optimized RLHF techniques, specifically designed for high-quality dialogue generation and role-playing scenarios.

Model Features

Multi-stage training

Combines supervised fine-tuning, QLoRA adapters, and GRPO-optimized RLHF techniques to ensure excellent model performance.

Role-playing optimization

Specially optimized for role-playing scenarios, supporting character card imports on platforms like SillyTavern.

Quantized version support

Offers multiple quantized versions (GGUF and Exl2) for easy deployment across different hardware environments.

Model Capabilities

Text generation

Dialogue systems

Role-playing

Context understanding

Use Cases

Entertainment

Role-playing chat

Engage in high-quality role-playing dialogues on platforms like SillyTavern.

Delivers natural and smooth character interaction experiences.

Creative writing

Story generation

Assists creators in generating storylines and dialogues.

Produces creative and engaging text content.

🚀 Captain-Eris_Violet Model

This project is a model developed using multi - stage supervised fine - tuning, pre - trained QLoRA adapters, and multi - stage RLHF optimized with GRPO. It provides various quantized versions and examples for SillyTavern.

🚀 Quick Start

The model image itself is now available as an importable character card for SillyTavern. This serves as an example of how to prepare your own card for use with this model. image/png

✨ Features

Developed using a combination of multi - stage supervised fine - tuning, pre - trained QLoRA adapters, and multi - stage RLHF optimized with GRPO.
The final model is created by merging the most promising candidates identified during the process.

📦 Installation

No specific installation steps are provided in the original document.

💻 Usage Examples

SillyTavern Reasoning Block Parsing Example

image/png

SillyTavern Mistral Formatting Example

Master Import Preset Here image/png

📚 Documentation

Training Notes

This model was developed using a combination of multi - stage supervised fine - tuning, pre - trained QLoRA adapters, and multi - stage RLHF optimized with GRPO. The final model was created by merging the most promising candidates identified during the process.

Series Comparison

image/png

Model Configuration

The following YAML configuration was used to produce this final version of the model:

slices:
  - sources:
      - model: Nitral-AI/Captain-Eris_Violet-0.420-Rebased
        layer_range: [0, 40]
      - model: Nitral-AI/Captain-Eris_Violet-GRPO-Rebased
        layer_range: [0, 40]
merge_method: slerp
base_model: Nitral-AI/Captain-Eris_Violet-0.420-Rebased
parameters:
  t:
    - filter: self_attn
      value: [0, 0.5, 0.3, 0.7, 1]
    - filter: mlp
      value: [1, 0.5, 0.7, 0.3, 0]
    - value: 0.420
dtype: bfloat16

Quants

Thanks to Mradermacher <3 Regular GGUF Imatrix GGUF 4bpw Exl2

📄 License

The license of this model is other.

📄 Information Table

Property	Details
Base Model	Nitral-AI/Captain-Eris_Violet-0.420-Rebased, Nitral-AI/Captain-Eris_Violet-GRPO-Rebased
Library Name	transformers
Tags	merge, finetune, GRPO, QLORA, SFT
License	other
Language	en

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご