This is a merge of pre-trained language models created using
mergekit. Experiment of merging top 3 7B models on the OpenLLm leaderboard (as of 5/30/2024)
This model was merged using the
DARE TIES merge method using
BarraHome/Mistroll-7B-v2.2 as a base.
1models:
2 - model: BarraHome/Mistroll-7B-v2.2
3 # no parameters necessary for base model
4 - model: yam-peleg/Experiment26-7B
5 parameters:
6 weight: 0.4
7 density: 0.7
8 - model: MTSAIR/multi_verse_model
9 parameters:
10 weight: 0.6
11 density: 0.7
12merge_method: dare_ties
13base_model: BarraHome/Mistroll-7B-v2.2
14parameters:
15 int8_mask: true
16dtype: bfloat16
