README.md

metadata

base_model:
  - Undi95/Unholy-8B-DPO-OAS
  - Undi95/Llama3-Unholy-8B-OAS
library_name: transformers
tags:
  - mergekit
  - merge

sin

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

Merge Details

Merge Method

This model was merged using the linear DARE merge method using Undi95/Unholy-8B-DPO-OAS as a base.

Models Merged

The following models were included in the merge:

• Undi95/Llama3-Unholy-8B-OAS
• merge/sin.a

Configuration

The following YAML configuration was used to produce this model:

base_model: Undi95/Unholy-8B-DPO-OAS
dtype: float32
merge_method: dare_linear
out_dtype: bfloat16
parameters:
  int8_mask: 1.0
  normalize: 0.0
slices:
- sources:
  - layer_range: [0, 32]
    model: Undi95/Llama3-Unholy-8B-OAS
    parameters:
      weight: 0.6
  - layer_range: [0, 32]
    model: merge/sin.a
    parameters:
      weight: 0.6
  - layer_range: [0, 32]
    model: Undi95/Unholy-8B-DPO-OAS
tokenizer_source: base