Resolving Interference When Merging Models
Paper
• 2306.01708 • Published
• 17
This is a merge of pre-trained language models created using mergekit.
This model was merged using the TIES merge method using win10/EVA-Meissa-mini-pro as a base.
The following models were included in the merge:
The following YAML configuration was used to produce this model:
models:
- model: shuttleai/shuttle-3-mini
parameters:
density: 0.2
weight: 0.2
- model: win10/EVA-Meissa-mini-pro
parameters:
density: 0.8
weight: 0.8
- model: shuttleai/shuttle-3-mini
parameters:
density: 0.2
weight: 0.2
- model: win10/EVA-Meissa-mini-pro
parameters:
density: 0.8
weight: 0.8
merge_method: ties
base_model: win10/EVA-Meissa-mini-pro
parameters:
normalize: true
int8_mask: true
dtype: bfloat16