bunnycore/Qwen-2.5-7B-R1-Stock
bunnycore/Qwen-2.5-7B-R1-Stock is a 7.6 billion parameter merged language model based on Qwen/Qwen2.5-7B-Instruct, created using the Model Stock method. It integrates components from DeepSeek-R1-Distill-Qwen-7B and a Qwen-2.5-7b-rp-lora adapter. This model is designed to combine the strengths of its constituent models, offering a broad utility for various language generation tasks with a 32K context length.
Loading preview...
Overview
bunnycore/Qwen-2.5-7B-R1-Stock is a 7.6 billion parameter language model created by bunnycore through a merge of several pre-trained models. This model leverages the Model Stock merge method, as described in the arXiv paper 2403.19522, to combine the capabilities of its base components. The primary base model for this merge is Qwen/Qwen2.5-7B-Instruct, known for its strong instruction-following abilities.
Key Capabilities
- Merged Architecture: Integrates
deepseek-ai/DeepSeek-R1-Distill-Qwen-7Bandbunnycore/Qwen-2.5-7b-rp-lorawithQwen/Qwen2.5-7B-Instructas the base, aiming for enhanced performance across diverse tasks. - Instruction Following: Benefits from the instruction-tuned nature of its Qwen base, making it suitable for various prompt-based applications.
- Broad Utility: Designed to offer general-purpose language generation and understanding, suitable for a range of NLP tasks.
- 32K Context Length: Supports a substantial context window, allowing for processing longer inputs and generating more coherent, extended outputs.
Good for
- General Language Tasks: Ideal for applications requiring robust text generation, summarization, and question answering.
- Experimentation with Merged Models: Provides a practical example of the Model Stock merging technique, useful for researchers and developers exploring model fusion.
- Applications requiring Qwen-based performance: Suitable for users familiar with or requiring the characteristics of the Qwen 2.5 series, with added influences from DeepSeek and a specific LoRA adapter.