saishf/Merge-Mayhem-L3-V2.1
Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:8BQuant:FP8Ctx Length:8kLicense:cc-by-nc-4.0Architecture:Transformer0.0K Open Weights Warm

saishf/Merge-Mayhem-L3-V2.1 is an 8 billion parameter language model created by saishf using the Model Stock merge method. It is based on openlynn/Llama-3-Soliloquy-8B-v2 and integrates components from several Llama-3-8B-Instruct derivatives, including those from ResplendentAI. This model is designed to combine the strengths of its constituent models, offering a versatile foundation for various generative AI tasks.

Loading preview...

Model Overview

saishf/Merge-Mayhem-L3-V2.1 is an 8 billion parameter language model developed by saishf. It was created using the Model Stock merge method, a technique described in the paper Model Stock, which combines multiple pre-trained language models to leverage their individual strengths.

Merge Details

This model uses openlynn/Llama-3-Soliloquy-8B-v2 as its base. It integrates components from several Llama-3-8B-Instruct derivatives, specifically those fine-tuned by ResplendentAI. The merged models include:

  • meta-llama/Meta-Llama-3-8B-Instruct + ResplendentAI/Luna_Llama3
  • meta-llama/Meta-Llama-3-8B-Instruct + ResplendentAI/RP_Format_QuoteAsterisk_Llama3
  • meta-llama/Meta-Llama-3-8B-Instruct + ResplendentAI/Smarts_Llama3
  • meta-llama/Meta-Llama-3-8B-Instruct + ResplendentAI/BlueMoon_Llama3
  • meta-llama/Meta-Llama-3-8B-Instruct + ResplendentAI/Aura_Llama3

Purpose

The merge aims to combine the diverse capabilities present in these specialized Llama-3 variants, potentially enhancing overall performance and versatility for generative tasks. The creator notes that this is a rerun of a previous version, with an updated configuration and base model, to explore potential improvements in the merged outcome.

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