llama3-8b-enhanced / README.md
mbahrsnc's picture
Upload folder using huggingface_hub
4a6d398 verified
|
raw
history blame
1.2 kB
metadata
base_model:
  - nvidia/Llama3-ChatQA-1.5-8B
  - failspy/Meta-Llama-3-8B-Instruct-abliterated-v3
library_name: transformers
tags:
  - mergekit
  - merge

merge

This is a merge of pre-trained language models 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:

Configuration

The following YAML configuration was used to produce this model:

slices:
  - sources:
      - model: failspy/Meta-Llama-3-8B-Instruct-abliterated-v3
        layer_range: [0, 32]
        output_weight: 0.3
      - model: nvidia/Llama3-ChatQA-1.5-8B
        layer_range: [0, 32]
        output_weight: 0.7
merge_method: slerp
base_model: nvidia/Llama3-ChatQA-1.5-8B
parameters:
  t:
    - filter: self_attn
      value: [0, 0.7, 0.5, 0.9, 1]
    - filter: mlp
      value: [1, 0.7, 0.9, 0.5, 0.3]
    - value: 0.7
dtype: bfloat16