Model soups: averaging weights of multiple fine-tuned models improves accuracy without increasing inference time
Paper
• 2203.05482 • Published
• 8
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:
model_name: "pre-cursa-o1-v1.6"
models:
- model: marcuscedricridia/cursa-o1-7b
parameters:
weight: 1.0
- model: marcuscedricridia/absolute-o1-7b
parameters:
weight: 1.0
- model: marcuscedricridia/cursa-o1-7b
parameters:
weight: 1.0
merge_method: linear
normalize: false
int8_mask: true
dtype: bfloat16
tokenizer_source: "union" # or "base" or a model path
chat_template: "auto" # or a template name or Jinja2 template