Editing Models with Task Arithmetic
Paper • 2212.04089 • Published • 8
How to use mergekit-community/mergekit-task_arithmetic-rycjsfm with Transformers:
# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("mergekit-community/mergekit-task_arithmetic-rycjsfm", dtype="auto")This is a merge of pre-trained language models created using mergekit.
This model was merged using the task arithmetic merge method using TouchNight/gemma-2-9b-it-untied as a base.
The following models were included in the merge:
The following YAML configuration was used to produce this model:
base_model: TouchNight/gemma-2-9b-it-untied
dtype: float32
merge_method: task_arithmetic
slices:
- sources:
- layer_range: [0, 42]
model: sam-paech/Darkest-muse-v1
parameters:
weight: 1.0
- layer_range: [0, 42]
model: IlyaGusev/gemma-2-9b-it-abliterated
parameters:
weight: 1.0
- layer_range: [0, 42]
model: TouchNight/gemma-2-9b-it-untied