intellectlabs/Kepler-8B-Instruct
Kepler-8B-Instruct is a 7.6 billion parameter instruction-tuned language model created by intellectlabs, merged using the DARE TIES method with Qwen/Qwen2.5-7B-Instruct as its base. It integrates capabilities from DeepSeek-R1-Distill-Qwen-7B, offering a 32768-token context length. This model is designed for general instruction-following tasks, leveraging its merged architecture for enhanced performance.
Loading preview...
Overview
Kepler-8B-Instruct is a 7.6 billion parameter instruction-tuned language model developed by intellectlabs. It was created using the DARE TIES merge method, combining the strengths of existing pre-trained models. The base model for this merge was Qwen/Qwen2.5-7B-Instruct, and it incorporated deepseek-ai/DeepSeek-R1-Distill-Qwen-7B to enhance its capabilities.
Key Characteristics
- Architecture: Merged model based on Qwen2.5-7B-Instruct.
- Parameter Count: 7.6 billion parameters.
- Context Length: Supports a context window of 32768 tokens.
- Merge Method: Utilizes the DARE TIES technique for combining model weights, aiming for improved performance and efficiency.
Use Cases
This model is suitable for a variety of instruction-following applications, benefiting from the combined knowledge and reasoning abilities of its constituent models. Its substantial context length makes it effective for tasks requiring processing longer inputs or generating extended responses.