fa-dorna-llama3-8b-instruct-x-meta-llama-3-8b-instruct-skt

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

Merge Details

Merge Method

This model was merged using the Spectral Knowledge Transfer (SKT) merge method using meta-llama/Meta-Llama-3-8B-Instruct as a base.

Models Merged

The following models were included in the merge:

• PartAI/Dorna-Llama3-8B-Instruct

Configuration

The following YAML configuration was used to produce this model:

dtype: bfloat16
tokenizer:
  source: union
merge_method: skt
models:
- model: PartAI/Dorna-Llama3-8B-Instruct
- model: meta-llama/Meta-Llama-3-8B-Instruct
base_model: meta-llama/Meta-Llama-3-8B-Instruct
parameters:
  beta: 12.0
  gamma: 1.0
  eps: 1.0e-08
  energy_keep: 0.98
  svd_on_cpu: false
  t_fallback: 0.5
write_readme: README.md