DELLA-Merging: Reducing Interference in Model Merging through Magnitude-Based Sampling
Paper • 2406.11617 • Published • 10
This is a merge of pre-trained language models created using mergekit.
This model was merged using the DELLA merge method using DreadPoor/Irix-12B-Model_Stock as a base.
The following models were included in the merge:
The following YAML configuration was used to produce this model:
base_model: DreadPoor/Irix-12B-Model_Stock
merge_method: della
models:
- model: mergekit-community/mergekit-dare_ties-twfgema
parameters:
weight: 0.5
- model: mergekit-community/mergekit-dare_ties-afgxxsc
parameters:
weight: 0.5
parameters:
density: 0.8325
normalize: true
tokenizer_source: union
chat_template: "chatml"
dtype: bfloat16