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 liminerity/Phigments12 as a base.
The following models were included in the merge:
The following YAML configuration was used to produce this model:
models:
- model: liminerity/Phigments12
# No parameters necessary for base model
- model: netcat420/MFANN3bv0.18
parameters:
density: 0.53
weight: 0.5
- model: netcat420/MFANN3bv0.6
parameters:
density: 0.53
weight: 0.3
- model: netcat420/MFANN3bv0.17.11
parameters:
density: 0.53
weight: 0.2
merge_method: dare_ties
base_model: liminerity/Phigments12
parameters:
int8_mask: true
dtype: bfloat16