phantasma

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

Merge Details

Merge Method

This model was merged using the NuSLERP merge method.

Models Merged

The following models were included in the merge:

• merge/spectre
• merge/twilight

Configuration

The following YAML configuration was used to produce this model:

chat_template: llama3
dtype: float32
merge_method: nuslerp
modules:
  default:
    slices:
    - sources:
      - layer_range: [0, 50]
        model: merge/twilight
        parameters:
          weight:
          - filter: self_attn
            value: [1.0, 0.7, 0.5, 0.3, 0.0]
          - filter: mlp
            value: [0.0, 0.3, 0.5, 0.7, 1.0]
          - value: 0.5
      - layer_range: [0, 50]
        model: merge/spectre
        parameters:
          weight:
          - filter: self_attn
            value: [0.0, 0.3, 0.5, 0.7, 1.0]
          - filter: mlp
            value: [1.0, 0.7, 0.5, 0.3, 0.0]
          - value: 0.5
tokenizer:
  pad_to_multiple_of: 32