metadata

base_model: []
library_name: transformers
tags:
  - mergekit
  - merge

Llama-3-Smolphin-8b

This is a merge of pre-trained language models abacusai/Llama-3-Smaug-8B and cognitivecomputations/dolphin-2.9-llama3-8b created using mergekit.

Merge Details

Merge Method

This model was merged using the SLERP merge method.

Models Merged

The following models were included in the merge:

abacusai/Llama-3-Smaug-8B
cognitivecomputations_dolphin-2.9-llama3-8b

Configuration

The following YAML configuration was used to produce this model:

slices:
  - sources:
      - model: cognitivecomputations/dolphin-2.9-llama3-8b
        layer_range: [0, 32]
      - model: abacusai/Llama-3-Smaug-8B
        layer_range: [0, 32]
merge_method: slerp
base_model: cognitivecomputations/dolphin-2.9-llama3-8b
parameters:
  t:
    - filter: self_attn
      value: [0, 0.5, 0.3, 0.7, 1]
    - filter: mlp
      value: [1, 0.5, 0.7, 0.3, 0]
    - value: 0.5
dtype: bfloat16