README.md

metadata

base_model:
  - mistralai/Mistral-7B-Instruct-v0.2
  - Nexusflow/Starling-LM-7B-beta
library_name: transformers
tags:
  - mergekit
  - merge
  - not-for-all-audiences

Mistralv0.2-StarlingLM-FrankenMaid-10.5B

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

Merge Details

See Below

Merge Method

This model was merged using the passthrough merge method and also dare_ties method.

Models Merged

The following models were included in the merge:

• mistralai/Mistral-7B-Instruct-v0.2
• Nexusflow/Starling-LM-7B-beta
• fM-v2

Configuration

Works well with Alpaca Instruct Template

The following YAML configuration was used to produce this model:

slices:
  - sources:
    - model: "fM-v2"
      layer_range: [0, 16]
  - sources:
    - model: "Nexusflow/Starling-LM-7B-beta"
      layer_range: [8, 24]
  - sources:
    - model: "mistralai/Mistral-7B-Instruct-v0.2"
      layer_range: [17, 32]
merge_method: passthrough
dtype: float16
name: fM-v2

models:
  - model: mistralai/Mistral-7B-Instruct-v0.2
  - model: Franken-Maid
    parameters:
      weight: 0.25
      density: 1.0
  - model: Franken-Maid
    parameters:
      weight: 0.30
      density: 1.0
merge_method: dare_ties
base_model: mistralai/Mistral-7B-Instruct-v0.2
parameters:
  int8_mask: true
dtype: bfloat16
name: fM-v2