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:
models:
- model: /local/scratch/zli2255/workspace/MergeExpert/checkpoints/baselines_openrs/len_MRL4096_ROLLOUT4_LR5e-7/global_step_30/actor/huggingface
parameters:
weight: 0.5
- model: /local/scratch/zli2255/workspace/MergeExpert/checkpoints/baselines_openrs/accfmt_MRL4096_ROLLOUT4_LR5e-7/global_step_54/actor/huggingface
parameters:
weight: 0.5
merge_method: linear
parameters:
normalize: true
dtype: bfloat16