DELLA-Merging: Reducing Interference in Model Merging through Magnitude-Based Sampling
Paper
•
2406.11617
•
Published
•
8
This is a merge of pre-trained language models created using mergekit.
This model was merged using the Linear DELLA merge method using mlabonne/Qwen3-4B-abliterated as a base.
The following models were included in the merge:
The following YAML configuration was used to produce this model:
merge_method: della_linear
parameters:
normalize: true
base_model: "mlabonne/Qwen3-4B-abliterated"
models:
- model: "POLARIS-Project/Polaris-4B-Preview"
parameters:
weight: 1
- model: "Menlo/Jan-nano-128k"
parameters:
weight: 1
- model: "ValiantLabs/Qwen3-4B-Esper3"
parameters:
weight: 1