README.md

metadata

base_model:
  - TheDrummer/Cydonia-24B-v2
  - Undi95/MistralThinker-v1.1
  - ReadyArt/Forgotten-Safeword-24B-V3.0
library_name: transformers
tags:
  - mergekit
  - merge

Cydonia-ForgottenThoughts-24B

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

Merge Details

Merge Method

This model was merged using the TIES merge method using TheDrummer/Cydonia-24B-v2 as a base.

Models Merged

The following models were included in the merge:

• Undi95/MistralThinker-v1.1
• ReadyArt/Forgotten-Safeword-24B-v3.0

Configuration

The following YAML configuration was used to produce this model:

models:
  # Base model
  - model: TheDrummer/Cydonia-24B-v2
  # Merge models
  - model: ReadyArt/Forgotten-Safeword-24B-v3.0
    parameters:
      density:
        - filter: self_attn
          value: 0.75
        - value: 0.5
      weight:
        - filter: self_attn
          value: [0.5, 0.35, 0.5, 0.35, 0.5]
        - filter: mlp
          value: [0.25, 0.35, 0.25, 0.35, 0.25]
        - value: 0.35
      lambda:
        - filter: self_attn
          value: 1.05
        - value: 0.9
  - model: Undi95/MistralThinker-v1.1
    parameters:
      density: 
        - filter: mlp
          value: 0.75
        - value: 0.5
      weight:
        - filter: mlp
          value: [0.5, 0.35, 0.5, 0.35, 0.5]
        - filter: self_attn
          value: [0.25, 0.35, 0.25, 0.35, 0.25]
        - value: 0.35
      lambda:
        - filter: mlp
          value: 1.05
        - value: 0.9
  
merge_method: ties
base_model: TheDrummer/Cydonia-24B-v2
parameters:
  normalize: true
dtype: float32