Language Models are Super Mario: Absorbing Abilities from Homologous Models as a Free Lunch
Paper
•
2311.03099
•
Published
•
30
This is a merge of pre-trained language models created using mergekit.
This model was merged using the DARE TIES merge method using anthracite-org/magnum-v2.5-12b-kto as a base.
The following models were included in the merge:
The following YAML configuration was used to produce this model:
models:
- model: anthracite-org/magnum-v2.5-12b-kto
#no parameters necessary for base model
- model: VAGOsolutions/SauerkrautLM-Nemo-12b-Instruct
parameters:
density: 0.5
weight: 0.5
- model: NeverSleep/Lumimaid-v0.2-12B
parameters:
density: 0.5
weight: 0.5
merge_method: dare_ties
base_model: anthracite-org/magnum-v2.5-12b-kto
parameters:
normalize: false
int8_mask: true
dtype: float16