Overview

This model, mergekit-ties-dhpqgnv, is a 1 billion parameter language model developed by Hjgugugjhuhjggg. It was created using the TIES merge method with meta-llama/Llama-3.2-1B as its base architecture. The TIES method combines the strengths of multiple fine-tuned models into a single, more generalized model.

Key Capabilities

This merged model integrates diverse functionalities from its constituent models, including:

• Mathematical and Reasoning Tasks: Incorporates models fine-tuned for MGSM8K, GSM8K, and general math/code problems.
• Medical Domain Knowledge: Includes components specialized in medical question answering (MedMCQA) and medical text analysis.
• SQL Generation: Features capabilities for generating SQL queries, including bilingual support.
• Hinglish Language Support: Contains a model fine-tuned on a Hinglish medical dataset.
• Common Sense Reasoning: Benefits from a model trained on common sense question answering.

Good For

This model is suitable for developers seeking a compact, multi-purpose LLM that can handle a variety of specialized tasks. Its merged nature makes it a versatile option for applications requiring a blend of mathematical, medical, coding, and general reasoning abilities within a 1 billion parameter footprint.