Gunulhona/Mistral-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 della merge method using Edentns/DataVortexM-7B-Instruct-v0.1 as a base.

Models Merged

The following models were included in the merge:

• Alphacode-AI/AlphaMist7B-slr-v3
• refarde/Mistral-7B-Instruct-v0.2-Ko-S-Core
• AIdenU/Mistral-7b-ko-Y24-DPO_v0.1

Configuration

The following YAML configuration was used to produce this model:

slices:
  - sources:
      - model: Edentns/DataVortexM-7B-Instruct-v0.1
        layer_range: [0, 32]
        parameters:
            weight: 1.0
      - model: refarde/Mistral-7B-Instruct-v0.2-Ko-S-Core
        layer_range: [0, 32]
        parameters:
            weight: 1.0
      - model: Alphacode-AI/AlphaMist7B-slr-v3
        layer_range: [0, 32]
        parameters:
            weight: 1.0
      - model: AIdenU/Mistral-7b-ko-Y24-DPO_v0.1
        layer_range: [0, 32]
        parameters:
            weight: 1.0
merge_method: della
base_model: Edentns/DataVortexM-7B-Instruct-v0.1
parameters:
  normalize: true
  int8_mask: true
  density: 0.7
  lambda: 1.1
  epsilon: 0.2
dtype: bfloat16