1dtype: bfloat16
2tokenizer_source: base
3merge_method: della_linear
4parameters:
5 density: 0.5
6base_model: migtissera/Tess-v2.5.2-Qwen2-72B
7models:
8 - model: Nexusflow/Athene-V2-Chat
9 parameters:
10 weight:
11 - filter: v_proj
12 value: [0, 0, 1, 1, 1, 1, 1, 1, 1, 0, 0]
13 - filter: o_proj
14 value: [1, 0, 1, 0, 0, 0, 0, 0, 1, 1, 1]
15 - filter: up_proj
16 value: [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
17 - filter: gate_proj
18 value: [0, 0, 1, 1, 1, 1, 1, 1, 1, 0, 0]
19 - filter: down_proj
20 value: [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]
21 - value: 0
22 - model: migtissera/Tess-v2.5.2-Qwen2-72B
23 parameters:
24 weight:
25 - filter: v_proj
26 value: [1, 1, 0, 0, 0, 0, 0, 0, 0, 1, 1]
27 - filter: o_proj
28 value: [0, 1, 0, 1, 1, 1, 1, 1, 0, 0, 0]
29 - filter: up_proj
30 value: [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]
31 - filter: gate_proj
32 value: [1, 1, 0, 0, 0, 0, 0, 0, 0, 1, 1]
33 - filter: down_proj
34 value: [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
35 - value: 1