ClaudioItaly/VoloGutenberg-9B

merge

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 SillyTilly/google-gemma-2-9b-it as a base.

Models Merged

The following models were included in the merge:

• inflatebot/G2-9B-Blackout-R1
• anthracite-org/magnum-v3-9b-customgemma2
• nbeerbower/Gemma2-Gutenberg-Doppel-9B
• lemon07r/Gemma-2-Ataraxy-9B
• sam-paech/Delirium-v1

Configuration

The following YAML configuration was used to produce this model:

models:
  - model: sam-paech/Delirium-v1  # Modello principale
    parameters:
      weight: 0.5  # Peso maggiore per il dataset Gutenberg
      density: 0.6  # Mantiene il 60% dei suoi parametri unici
  - model: nbeerbower/Gemma2-Gutenberg-Doppel-9B
    parameters:
      weight: 0.2
      density: 0.4
  - model: lemon07r/Gemma-2-Ataraxy-9B
    parameters:
      weight: 0.15
      density: 0.3
  - model: inflatebot/G2-9B-Blackout-R1
    parameters:
      weight: 0.1
      density: 0.3
  - model: anthracite-org/magnum-v3-9b-customgemma2
    parameters:
      weight: 0.05
      density: 0.2

merge_method: ties
base_model: SillyTilly/google-gemma-2-9b-it
tokenizer_source: sam-paech/Delirium-v1  # Usa il tokenizer del modello principale
dtype: float16

parameters:
  normalize: true
  alpha: 0.35  # Bilancia tra task vectors e base model