Resolving Interference When Merging Models
Paper
•
2306.01708
•
Published
•
15
This is a merge of pre-trained language models created using mergekit.
This model was merged using the TIES merge method using fatafix0000/aaa0 as a base.
The following models were included in the merge:
The following YAML configuration was used to produce this model:
base_model: fatafix0000/aaa0
dtype: float16
merge_method: ties
slices:
- sources:
- layer_range: [0, 40]
model: fatafix0000/aaa0
parameters:
density: [1.0, 0.7, 0.1]
weight: 1.0
- layer_range: [0, 40]
model: fatafix0000/aaa3
parameters:
density: 0.3
weight: 0.0001