README.md

metadata

base_model:
  - gradientai/Llama-3-8B-Instruct-262k
  - Nitral-AI/Echidna-7b-128k
library_name: transformers
tags:
  - mergekit
  - merge

merge

This is a merge of pre-trained language models created using mergekit.

Merge Details

Merge Method

This model was merged using the DARE TIES merge method using Nitral-AI/Echidna-7b-128k as a base.

Models Merged

The following models were included in the merge:

• gradientai/Llama-3-8B-Instruct-262k

Configuration

The following YAML configuration was used to produce this model:

models:
  - model: Nitral-AI/Echidna-7b-128k
    #no parameters necessary for base model
  - model: gradientai/Llama-3-8B-Instruct-262k
    parameters:
      density: 0.5
      weight: 0.5

merge_method: dare_ties
base_model: Nitral-AI/Echidna-7b-128k
parameters:
  normalize: false
  int8_mask: true
dtype: float16