mergekit-community/mergekit-dare_ties-yhoaczs

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 meta-llama/Meta-Llama-3-8B-Instruct as a base.

Models Merged

The following models were included in the merge:

• Columbia-NLP/LION-LLaMA-3-8b-sft-v1.0
• akjindal53244/Llama-3.1-Storm-8B
• mlabonne/OrpoLlama-3-8B

Configuration

The following YAML configuration was used to produce this model:

models:
  - model: meta-llama/Meta-Llama-3-8B-Instruct
  - model: akjindal53244/Llama-3.1-Storm-8B
    parameters:
      density: 0.53
      weight: 0.45
  - model: Columbia-NLP/LION-LLaMA-3-8b-sft-v1.0
    parameters:
      density: 0.5
      weight: 0.25
  - model: mlabonne/OrpoLlama-3-8B
    parameters:
      density: 0.5
      weight: 0.15

merge_method: dare_ties
base_model: meta-llama/Meta-Llama-3-8B-Instruct
parameters:
  normalize: true
  int8_mask: true
dtype: bfloat16