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 Undi95/PsyMedRP-v1-20B as a base.
The following models were included in the merge:
The following YAML configuration was used to produce this model:
# Testing random seeding on a 3 solid models.
models:
- model: Undi95/PsyMedRP-v1-20B
parameters:
weight: 0.6
density: 0.8
- model: Undi95/MXLewd-L2-20B
parameters:
weight: 0.6
density: 0.8
- model: Elfrino/PsyMedRose-20B
parameters:
weight: 0.38
density: 0.6
merge_method: dare_ties
base_model: Undi95/PsyMedRP-v1-20B
random_seed: 123
dtype: bfloat16