kaitchup/Mayonnaise-4in1-03

TEXT GENERATIONConcurrency Cost:1Model Size:7BQuant:FP8Ctx Length:8kPublished:Jan 27, 2024License:apache-2.0Architecture:Transformer0.0K Open Weights Cold

kaitchup/Mayonnaise-4in1-03 is a 7 billion parameter causal language model developed by The Kaitchup, built using mergekit based on mistralai/Mistral-7B-v0.1. This model is a mixture of experts (MoE) created with a TIES-merging method, combining mncai/mistral-7b-dpo-v5, flemmingmiguel/MBX-7B, and BarryFutureman/NeuralTurdusVariant1-7B. It is designed for general English language tasks, leveraging its merged architecture to potentially enhance performance over its base models.

Loading preview...

Overview

kaitchup/Mayonnaise-4in1-03 is a 7 billion parameter causal language model developed by The Kaitchup. It is an English-language model built upon the mistralai/Mistral-7B-v0.1 architecture, utilizing a Mixture of Experts (MoE) approach. The model was constructed using the mergekit tool, specifically employing the TIES-merging method.

Key Capabilities

This model is a composite of several fine-tuned models, merged to potentially leverage their individual strengths. The merging configuration includes:

  • Base Model: mncai/mistral-7b-dpo-v5
  • Merged Experts: flemmingmiguel/MBX-7B and BarryFutureman/NeuralTurdusVariant1-7B, with specified densities and weights.

Unique Approach

The creation of Mayonnaise-4in1-03 is detailed in the article "The Mayonnaise: Rank First on the Open LLM Leaderboard with TIES-Merging", which outlines the recipe for achieving high performance through TIES-merging. This method aims to combine the knowledge and capabilities of multiple models into a single, more robust model. The model is licensed under Apache 2.0.

Popular Sampler Settings

Top 3 parameter combinations used by Featherless users for this model. Click a tab to see each config.

temperature
top_p
top_k
frequency_penalty
presence_penalty
repetition_penalty
min_p