Model soups: averaging weights of multiple fine-tuned models improves accuracy without increasing inference time
Paper
•
2203.05482
•
Published
•
7
This is a merge of pre-trained language models created using mergekit.
This model was merged using the Linear merge method.
The following models were included in the merge:
The following YAML configuration was used to produce this model:
models:
- model: FPHam/L3-8B-Everything-COT+ResplendentAI/NoWarning_Llama3
- model: FPHam/L3-8B-Everything-COT+Azazelle/Llama-3-LongStory-LORA
- model: FPHam/L3-8B-Everything-COT+Azazelle/Llama-3-8B-Abomination-LORA
- model: FPHam/L3-8B-Everything-COT+nothingiisreal/llama3-8B-DWP-lora
parameters:
weight: 1.0
merge_method: linear
dtype: float16