Luminea-Dare
This is a merge of pre-trained language models created using mergekit. Seems to be better than the model stock version, retains nvidia/Llama-3.1-Nemotron-70B-Instruct-HF style formatting, while writing like NeverSleep/Lumimaid-v0.2-70B
Merge Details
Merge Method
This model was merged using the DARE TIES merge method using Arkhiveus/L3.1-70B-Luminea as a base.
Models Merged
The following models were included in the merge:
Configuration
The following YAML configuration was used to produce this model:
models:
- model: nvidia/Llama-3.1-Nemotron-70B-Instruct-HF
parameters:
weight: 1.0
density: 1.0
- model: NeverSleep/Lumimaid-v0.2-70B
parameters:
weight: 1.0
density: 1.0
base_model: Arkhiveus/L3.1-70B-Luminea
merge_method: dare_ties
dtype: float16
name: Luminea-Dare