bunnycore/Qwen-2.5-7B-Deep-Sky-T1

TEXT GENERATIONConcurrency Cost:1Model Size:7.6BQuant:FP8Ctx Length:32kPublished:Feb 16, 2025Architecture:Transformer Cold

bunnycore/Qwen-2.5-7B-Deep-Sky-T1 is a 7.6 billion parameter language model created by bunnycore, merged using the TIES method from a Qwen2.5-7B-RRP-1M base. This model integrates components from NovaSky-AI/Sky-T1-mini and bunnycore/Qwen-2.5-7B-Deep-Stock-v5, offering a combined set of capabilities. It is designed for general language tasks, leveraging its merged architecture to potentially enhance performance across various applications.

Loading preview...

Model Overview

bunnycore/Qwen-2.5-7B-Deep-Sky-T1 is a 7.6 billion parameter language model developed by bunnycore. It was created using the TIES merge method, which combines the strengths of multiple pre-trained models into a single, more capable model. The base model for this merge was bunnycore/Qwen2.5-7B-RRP-1M.

Merge Details

This model is a composite of two distinct models, carefully weighted to achieve its final characteristics:

  • NovaSky-AI/Sky-T1-mini: Contributed with a density of 0.3 and a weight of 0.3.
  • bunnycore/Qwen-2.5-7B-Deep-Stock-v5: Contributed with a density of 0.5 and a weight of 0.5.

The merge process utilized mergekit and was configured with int8_mask: true and dtype: float16, indicating an optimization for efficiency while maintaining performance. This approach allows the model to inherit and blend the specific capabilities and knowledge bases of its constituent models.

Potential Use Cases

Given its merged nature, bunnycore/Qwen-2.5-7B-Deep-Sky-T1 is likely suitable for a broad range of general-purpose language generation and understanding tasks, potentially excelling in areas where its merged components individually showed strength.