This is a merge of pre-trained language models created using
mergekit.
This model was merged using the
TIES merge method using
Qwen/Qwen2.5-0.5B-Instruct as a base.
1slices:
2 - sources:
3 - model: Qwen/Qwen2.5-0.5B-Instruct
4 layer_range: [0, 11]
5 - model: Qwen/Qwen2.5-0.5B-Instruct
6 layer_range: [12, 23]
7merge_method: ties
8base_model: Qwen/Qwen2.5-0.5B-Instruct
9parameters:
10 t:
11 - filter: self_attn
12 value: [0.1, 0.3, 0.5, 0.7, 0.9]
13 - filter: mlp
14 value: [0.1, 0.3, 0.5, 0.7, 0.9]
15 - value: 0.5
16dtype: bfloat16
17
