vishnukv/newmerge

TEXT GENERATIONConcurrency Cost:1Model Size:7BQuant:FP8Ctx Length:4kPublished:Mar 5, 2024License:mitArchitecture:Transformer Open Weights Cold

vishnukv/newmerge is a 7 billion parameter language model created by vishnukv, resulting from a merge of pre-trained models using the DARE TIES method. It is based on uukuguy/speechless-mistral-dolphin-orca-platypus-samantha-7b and incorporates PetroGPT/WestSeverus-7B-DPO and vishnukv/WestSeverusJaskier. This model is designed to combine the strengths of its constituent models for general language generation tasks.

Loading preview...

Overview

vishnukv/newmerge is a 7 billion parameter language model developed by vishnukv. It was created using the mergekit tool, specifically employing the DARE TIES merge method.

Merge Method

The model's architecture is based on a merge of several pre-trained language models. The merging process utilized the DARE (Direct Acyclic Graph Ensemble) and TIES (Task-Independent Ensemble of Specialists) methods, which are designed to combine the capabilities of multiple models effectively. The base model for this merge was uukuguy/speechless-mistral-dolphin-orca-platypus-samantha-7b.

Constituent Models

The newmerge model integrates knowledge and capabilities from the following specific models:

Key Capabilities

  • Combines features from multiple specialized 7B models.
  • Leverages the DARE TIES merging technique for potentially enhanced performance across various tasks.

Good For

  • General language generation and understanding tasks.
  • Experimentation with merged model architectures.