Note: A much-improved version is available at
jsgreenawalt/gemma-2-9B-it-advanced-v2.1Experimental merge #1, attempting to combine some of the advanced Gemma fine-tunes
This is a merge of pre-trained language models created using
mergekit.
This model was merged using the della merge method using
google/google-gemma-2-9b-it as a base.
1models:
2 - model: google/gemma-2-9b-it
3 # no parameters necessary for base model
4 - model: princeton-nlp/gemma-2-9b-it-SimPO
5 parameters:
6 density: 0.5
7 weight: 0.5
8 - model: wzhouad/gemma-2-9b-it-WPO-HB
9 parameters:
10 density: 0.5
11 weight: 0.5
12merge_method: della
13base_model: google/gemma-2-9b-it
14parameters:
15 normalize: true
16dtype: float16
17
