metadata

base_model:
  - IntervitensInc/Mistral-Nemo-Base-2407-chatml
library_name: transformers
tags:
  - mergekit
  - merge

Welcome, brave one; you've come a long mile.

magmell-r1g

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

Merge Details

Multi-stage SLERP merge, DARE-TIES'd together. Intended to be a general purpose "Best of Nemo" model for any fictional, creative use case. Inspired by hyper-merges like Tiefighter and Umbral Mind.

Hero (RP, kink/trope coverage): Chronos Gold, Sunrose. Monk (Intelligence, groundedness): Bophades, Wissenschaft. Deity (Prose, flair): Gutenberg, Magnum 2.5 KTO.

Merge Method

This model was merged using the DARE TIES merge method using IntervitensInc/Mistral-Nemo-Base-2407-chatml as a base.

Models Merged

The following models were included in the merge:

/workspace/models/magmell-r1/monk
/workspace/models/magmell-r1/hero
/workspace/models/magmell-r1/deity

Configuration

The following YAML configuration was used to produce this model:

base_model: IntervitensInc/Mistral-Nemo-Base-2407-chatml
merge_method: dare_ties
slices:
- sources:
  - layer_range: [0, 40]
    model: /workspace/models/magmell-r1/monk
    parameters:
      density: 0.7
      weight: 0.5
  - layer_range: [0, 40]
    model: /workspace/models/magmell-r1/hero
    parameters:
      density: 0.9
      weight: 1.0
  - layer_range: [0, 40]
    model: /workspace/models/magmell-r1/deity
    parameters:
      density: 0.5
      weight: 0.7
  - layer_range: [0, 40]
    model: IntervitensInc/Mistral-Nemo-Base-2407-chatml
tokenizer_source: base