yamatazen/Twilight-SCE-12B

This is a Mistral model with ChatML tokens added to the tokenizer.

merge

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

Merge Details

Merge Method

This model was merged using the SCE merge method using shisa-ai/shisa-v2-mistral-nemo-12b as a base.

Models Merged

The following models were included in the merge:

• inflatebot/MN-12B-Mag-Mell-R1
• yamatazen/Himeyuri-Magnum-12B
• Elizezen/Himeyuri-v0.1-12B

Configuration

The following YAML configuration was used to produce this model:

base_model: shisa-ai/shisa-v2-mistral-nemo-12b
models:
  - model: Elizezen/Himeyuri-v0.1-12B
    parameters:
      weight: 1.0
  - model: yamatazen/Himeyuri-Magnum-12B
    parameters:
      weight: 0.6
  - model: inflatebot/MN-12B-Mag-Mell-R1
    parameters:
      weight: 0.3
merge_method: sce
dtype: bfloat16
parameters:
  normalize: true
  select_topk: 0.5
tokenizer:
  source: union