mergekit-community/MN-Hekate-Nyktipolos-17B

merge

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

Merge Details

Merge Method

This model was merged using the Model Stock merge method using mergekit-community/MN-Hekate-Daidalos-17B as a base.

Models Merged

The following models were included in the merge:

• nbeerbower/Mistral-Nemo-Gutenberg-Doppel-12B
• mistralai/Mistral-Nemo-Base-2407
• nbeerbower/Lyra-Gutenberg-mistral-nemo-12B
• nbeerbower/mistral-nemo-bophades-12B
• HumanLLMs/Human-Like-Mistral-Nemo-Instruct-2407
• mergekit-community/MN-Sappho-g3-12B
• Nitral-AI/Captain-Eris_Violet-GRPO-v0.420

Configuration

The following YAML configuration was used to produce this model:

dtype: float32
out_dtype: bfloat16
merge_method: model_stock
base_model: mergekit-community/MN-Hekate-Daidalos-17B
slices:
  - sources:
    - model: mergekit-community/MN-Hekate-Daidalos-17B
      layer_range: [0, 20]

  - sources:
    - model: mergekit-community/MN-Hekate-Daidalos-17B
      layer_range: [20, 36]
      parameters:
        weight: 3
    - model: mistralai/Mistral-Nemo-Base-2407
      layer_range: [16, 32]
      parameters:
        weight: 2
    - model: Nitral-AI/Captain-Eris_Violet-GRPO-v0.420
      layer_range: [16, 32]
      parameters:
        weight: [1.5, 1.49, 1.46, 1.4, 1.33, 1.25, 1.15, 1.05, 1]
    - model: nbeerbower/mistral-nemo-bophades-12B
      layer_range: [16, 32]
      parameters:
        weight: [1.5, 1.49, 1.46, 1.4, 1.33, 1.25, 1.15, 1.05, 1]
    - model: mergekit-community/MN-Sappho-g3-12B
      layer_range: [16, 32]
      parameters:
        weight: [1.0, 1.1, 1.2, 1.29, 1.37, 1.43, 1.48, 1.5, 1.5]
    - model: nbeerbower/Lyra-Gutenberg-mistral-nemo-12B
      layer_range: [16, 32]
      parameters:
        weight: [1.0, 1.1, 1.2, 1.29, 1.37, 1.43, 1.48, 1.5, 1.5]

  - sources:
    - model: mergekit-community/MN-Hekate-Daidalos-17B
      layer_range: [36, 56]
      parameters:
        weight: 3
    - model: nbeerbower/Mistral-Nemo-Gutenberg-Doppel-12B
      layer_range: [20, 40]
      parameters:
        weight: [1, 0.75]
    - model: Nitral-AI/Captain-Eris_Violet-GRPO-v0.420
      layer_range: [20, 40]
      parameters:
        weight: [1, 0.5]
    - model: HumanLLMs/Human-Like-Mistral-Nemo-Instruct-2407
      layer_range: [20, 40]
      parameters:
        weight: [1, 0.75]

tokenizer:
  source: union
  tokens:
    "[INST]":
      source: mergekit-community/MN-Hekate-Daidalos-17B
      force: true
    "[/INST]":
      source: mergekit-community/MN-Hekate-Daidalos-17B
      force: true
    "<|im_start|>":
      source: mergekit-community/MN-Hekate-Daidalos-17B
      force: true
    "<|im_end|>":
      source: mergekit-community/MN-Hekate-Daidalos-17B
      force: true