Resolving Interference When Merging Models
Paper
•
2306.01708
•
Published
•
15
This is a merge of pre-trained language models created using mergekit.
Very first merge...
This model was merged using the TIES merge method using Magnum-Picaro-0.7-v2-12b as a base.
The following models were included in the merge:
The following YAML configuration was used to produce this model:
models:
- model: F:\AI\Pic\Magnum-Picaro-0.7-v2-12b
parameters:
density: [1, 0.7, 0.1] # density gradient
weight: 1.0
- model: F:\AI\Wayfarer\Wayfarer-12B
parameters:
density: 0.3
weight: [0, 0.3, 0.7, 1] # weight gradient
merge_method: ties
base_model: F:\AI\Pic\Magnum-Picaro-0.7-v2-12b
parameters:
normalize: true
int8_mask: true
dtype: float16