L3-8B-EnchantedForest-v0.5

merge

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

Notes

0.5 is much smarter compared to 0.6 to 0.8. It will serve as a checkpoint in training.

Merge Details

Merge Method

This model was merged using the SLERP merge method.

Models Merged

The following models were included in the merge:

• Rockinsince87/L3-8B-EnchantedForest-v0.4
• Dampfinchen/Llama-3-8B-Ultra-Instruct

Configuration

The following YAML configuration was used to produce this model:

slices:
- sources:
  - model: Rockinsince87/L3-8B-EnchantedForest-v0.4
    layer_range:
    - 0
    - 32
  - model: Dampfinchen/Llama-3-8B-Ultra-Instruct
    layer_range:
    - 0
    - 32
merge_method: slerp
base_model: Rockinsince87/L3-8B-EnchantedForest-v0.4
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