Gunulhona/Llama-Ko-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 Saxo/Linkbricks-Horizon-AI-Korean-llama-3.1-sft-dpo-8B as a base.

Models Merged

The following models were included in the merge:

• Saxo/Linkbricks-Horizon-AI-Korean-llama3-sft-dpo-8b-base
• tesser-ai/Tesser-Llama-3-Ko-8B
• NousResearch/Hermes-3-Llama-3.1-8B
• meta-llama/Meta-Llama-3.1-8B
• maum-ai/Llama-3-MAAL-8B-Instruct-v0.1

Configuration

The following YAML configuration was used to produce this model:

models:
    - model: tesser-ai/Tesser-Llama-3-Ko-8B
      layer_range: [0, 32]
      parameters:
          density: 0.5
          weight: 0.45
    - model: maum-ai/Llama-3-MAAL-8B-Instruct-v0.1
      layer_range: [0, 32]
      parameters:
          density: 0.5
          weight: 0.45
    - model: meta-llama/Meta-Llama-3.1-8B
      layer_range: [0, 32]
      parameters:
          density: 0.5
          weight: 0.45
    - model: NousResearch/Hermes-3-Llama-3.1-8B
      layer_range: [0, 32]
      parameters:
          density: 0.5
          weight: 0.45
    - model: Saxo/Linkbricks-Horizon-AI-Korean-llama3-sft-dpo-8b-base
      layer_range: [0, 32]
      parameters:
          density: 0.5
          weight: 0.45
    - model: Saxo/Linkbricks-Horizon-AI-Korean-llama-3.1-sft-dpo-8B
      layer_range: [0, 32]
      parameters:
          density: 0.5
          weight: 0.45
merge_method: dare_ties
base_model: Saxo/Linkbricks-Horizon-AI-Korean-llama-3.1-sft-dpo-8B
dtype: bfloat16