kenonix/kor-llama3.1

output_folder11

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 /run/media/kenonix/0054F25B7C8C7F23/make/output_folder10 as a base.

Models Merged

The following models were included in the merge:

• /run/media/kenonix/0054F25B7C8C7F23/llama3.1/Llama-3.1-Korean-8B-Instruct
• /run/media/kenonix/0054F25B7C8C7F23/llama3.1/Llama-VARCO-8B-Instruct
• /run/media/kenonix/0054F25B7C8C7F23/llama3.1/Linkbricks-Horizon-AI-Korean-Pro-8B
• /run/media/kenonix/0054F25B7C8C7F23/llama3.1/llama-3.1-8b-komedic-instruct

Configuration

The following YAML configuration was used to produce this model:

models:
  - model: /run/media/kenonix/0054F25B7C8C7F23/make/output_folder10
    # no parameters necessary for base model
  - model: /run/media/kenonix/0054F25B7C8C7F23/llama3.1/llama-3.1-8b-komedic-instruct
    parameters:
      density: 0.6
      weight: 0.5
  - model: /run/media/kenonix/0054F25B7C8C7F23/llama3.1/Linkbricks-Horizon-AI-Korean-Pro-8B
    parameters:
      density: 0.4
      weight: 0.2
  - model: /run/media/kenonix/0054F25B7C8C7F23/llama3.1/Llama-3.1-Korean-8B-Instruct
    parameters:
      density: 0.4
      weight: 0.5
  - model: /run/media/kenonix/0054F25B7C8C7F23/llama3.1/Llama-VARCO-8B-Instruct
    parameters:
      density: 0.4
      weight: 0.5

merge_method: dare_ties
base_model: /run/media/kenonix/0054F25B7C8C7F23/make/output_folder10
dtype: bfloat16