This is a merge of pre-trained language models created using
mergekit.
This model was merged using the
DARE TIES merge method using
Undi95/PsyMedRP-v1-20B as a base.
1# Testing random seeding on a 3 solid models.
2
3models:
4 - model: Undi95/PsyMedRP-v1-20B
5 parameters:
6 weight: 0.6
7 density: 0.8
8 - model: Undi95/MXLewd-L2-20B
9 parameters:
10 weight: 0.6
11 density: 0.8
12 - model: Elfrino/PsyMedRose-20B
13 parameters:
14 weight: 0.38
15 density: 0.6
16
17merge_method: dare_ties
18base_model: Undi95/PsyMedRP-v1-20B
19random_seed: 123
20dtype: bfloat16
