Resolving Interference When Merging Models
Paper
•
2306.01708
•
Published
•
15
This is a merge of pre-trained language models created using mergekit.
This model was merged using the TIES merge method using /GLOBALFS/gznwp_3/qxj/liangjh/model_set/qwen/Qwen2-Audio-7B-Instruct as a base.
The following models were included in the merge:
The following YAML configuration was used to produce this model:
models:
- model: /GLOBALFS/gznwp_3/qxj/liangjh/LLaMA-Factory-audio/output/Qwen2-Audio-7B-Instruct-train-all-full-new/checkpoint-4566
parameters:
density: 0.6
weight: 0.5
- model: /GLOBALFS/gznwp_3/qxj/liangjh/LLaMA-Factory-audio/output/Qwen2-Audio-7B-Instruct-train-all-full-new/checkpoint-9132
parameters:
density: 0.6
weight: 0.5
- model: /GLOBALFS/gznwp_3/qxj/liangjh/LLaMA-Factory-audio/output/Qwen2-Audio-7B-Instruct-train-all-full-new/checkpoint-13698
parameters:
density: 0.6
weight: 0.5
merge_method: ties
base_model: /GLOBALFS/gznwp_3/qxj/liangjh/model_set/qwen/Qwen2-Audio-7B-Instruct
parameters:
normalize: true
dtype: float16