arcee-ai/Biomistral-Calme-Instruct-7b

Biomistral-Calme-Instruct-7b Overview

Biomistral-Calme-Instruct-7b is a 7 billion parameter language model developed by arcee-ai. It is a merged model, combining the strengths of two distinct base models: MaziyarPanahi/Calme-7B-Instruct-v0.1.1 and BioMistral/BioMistral-7B.

Key Characteristics

• Hybrid Capabilities: This model integrates general instruction-following abilities from the Calme-7B-Instruct model with the specialized biomedical knowledge present in BioMistral-7B.
• Merge Method: The model was created using mergekit with a slerp merge method, carefully blending the weights of the constituent models across different layers.
• Parameter Configuration: Specific t values were applied to self-attention and MLP layers during the merge process to optimize the combination of features from both base models.

Potential Use Cases

• Biomedical Q&A: Answering questions that require both general understanding and specific knowledge in biology, medicine, or related scientific fields.
• Instruction Following: Executing complex instructions in contexts that might involve scientific terminology or concepts.
• Research Assistance: Aiding in tasks like summarizing scientific papers, generating hypotheses, or extracting information from biomedical texts, while maintaining conversational coherence.