Language Models are Super Mario: Absorbing Abilities from Homologous Models as a Free Lunch
Paper
• 2311.03099 • Published
• 30
This is a merge of pre-trained language models created using mergekit.
This model was merged using the DARE TIES merge method using unsloth/Qwen2.5-14B-Instruct as a base.
The following models were included in the merge:
The following YAML configuration was used to produce this model:
models:
- model: unsloth/Qwen2.5-14B-Instruct
parameters:
weight: 0.575
density: 0.575
- model: rombodawg/Rombos-LLM-V2.6-Qwen-14b
parameters:
weight: 0.425
density: 0.425
merge_method: dare_ties
base_model: unsloth/Qwen2.5-14B-Instruct
dtype: bfloat16