CyberpunkLegend/Qwen2.5-7B-Instruct-CharacterEnhance

TEXT GENERATIONConcurrent Unit Cost:1Model Size:7.6BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Jun 19, 2026License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Cold

CyberpunkLegend/Qwen2.5-7B-Instruct-CharacterEnhance is a 7.6 billion parameter QLoRA fine-tuned model based on Qwen2.5-7B-Instruct, specifically optimized for bilingual (English and Chinese) character role-play dialogue generation. It excels at producing natural, character-consistent responses in conversational AI scenarios. The model was trained on the PIPPA dataset to enhance its ability to maintain character personas.

Loading preview...

Model Overview

CyberpunkLegend/Qwen2.5-7B-Instruct-CharacterEnhance is a 7.6 billion parameter language model, fine-tuned using QLoRA (4-bit NF4 quantization) on the robust Qwen2.5-7B-Instruct base. Its primary focus is bilingual character role-play dialogue generation, supporting both English and Chinese.

Key Capabilities

  • Character-Consistent Dialogue: Generates natural and in-character responses for role-play scenarios.
  • Bilingual Support: Trained on a curated dataset including both English and Chinese dialogues from the PIPPA dataset.
  • Efficient Fine-tuning: Utilizes QLoRA with a LoRA Rank of 8, allowing for efficient adaptation.

Training Details

The model was trained for 1 epoch on 3,044 samples (1,522 English + 1,522 Chinese) with a maximum sequence length of 2048 tokens. The training process took approximately 45 minutes on an RTX 5080 16GB GPU, achieving a final evaluation loss of 1.9628.

Limitations

  • Inherits biases from the PIPPA dataset and the base Qwen2.5-7B-Instruct model.
  • Responses are typically constrained to around 30 Chinese characters (or equivalent), making it less suitable for long-form text generation.
  • Character consistency may decrease for personas or scenarios outside the training data distribution.

When to Use This Model

This model is ideal for applications requiring short, character-driven conversational AI, particularly in bilingual contexts. It is well-suited for interactive storytelling, virtual companions, or any scenario where maintaining a specific persona in dialogue is crucial.