This is a merge of pre-trained language models created using
mergekit.
This model was merged using the
TIES merge method using
mistralai/Mistral-7B-v0.1 as a base.
1base_model: mistralai/Mistral-7B-v0.1
2dtype: float16
3merge_method: ties
4modules:
5 default:
6 slices:
7 - sources:
8 - layer_range: [0, 32]
9 model: mistralai/Mistral-7B-v0.1
10 - layer_range: [0, 32]
11 model: OpenPipe/mistral-ft-optimized-1218
12 parameters:
13 density: 0.5
14 weight: 0.5
15 - layer_range: [0, 32]
16 model: mlabonne/NeuralHermes-2.5-Mistral-7B
17 parameters:
18 density: 0.5
19 weight: 0.3
20parameters:
21 normalize: 1.0
