metadata
base_model:
- senseable/WestLake-7B-v2
- NousResearch/Yarn-Mistral-7b-128k
tags:
- mergekit
- merge
license: apache-2.0
Merged
This is a merge of pre-trained language models created using mergekit, but using my experimental branch swapping
The "Other" branch is where I did something wrong, so in the "main" I did it right(I hope 😅).
Merge Details
Merge Method
This model was merged using the task_swapping merge method not sure if I did it right and how the model was impacted, using NousResearch/Yarn-Mistral-7b-128k as a base.
Models Merged
The following models were included in the merge:
Configuration
The following YAML configuration was used to produce this model:
base_model:
model:
path: NousResearch/Yarn-Mistral-7b-128k
dtype: bfloat16
merge_method: task_swapping
slices:
- sources:
- layer_range: [0, 32]
model:
model:
path: senseable/WestLake-7B-v2
parameters:
weight: 0.666
diagonal_offset: 2
invert_offset: True
- layer_range: [0, 32]
model:
model:
path: NousResearch/Yarn-Mistral-7b-128k