Resolving Interference When Merging Models
Paper
•
2306.01708
•
Published
•
15
This model was meant as a component model for a larger merge but is usable as is.
This is a merge of pre-trained language models created using mergekit.
This model was merged using the TIES merge method using IntervitensInc/Mistral-Nemo-Base-2407-chatml as a base.
The following models were included in the merge:
The following YAML configuration was used to produce this model:
models:
- model: IntervitensInc/Mistral-Nemo-Base-2407-chatml
- model: unsloth/Mistral-Nemo-Instruct-2407
parameters:
weight: [0.50, 0.20]
density: [0.75, 0.55]
- model: anthracite-org/magnum-12b-v2
parameters:
weight: [0.50, 0.80]
density: [0.75, 0.85]
merge_method: ties
base_model: IntervitensInc/Mistral-Nemo-Base-2407-chatml
parameters:
int8_mask: true
rescale: true
normalize: false
dtype: bfloat16