Views
No views yet
1models:
2 - model: TareksLab/M-MERGE4
3 parameters:
4 weight:
5 - filter: self_attn
6 value: [0.3, 0.1, 0.2]
7 - filter: mlp
8 value: [0.4, 0.2, 0.1]
9 - value: 0.2
10 density: 0.7
11 lambda: 1.05
12 - model: TareksLab/M-MERGE3
13 parameters:
14 weight:
15 - filter: self_attn
16 value: [0.2, 0.1, 0.3]
17 - filter: mlp
18 value: [0.3, 0.1, 0.2]
19 - value: 0.2
20 density: 0.65
21 lambda: 1.05
22 - model: TareksLab/M-MERGE2
23 parameters:
24 weight:
25 - filter: self_attn
26 value: [0.1, 0.3, 0.1]
27 - filter: mlp
28 value: [0.2, 0.3, 0.1]
29 - value: 0.2
30 density: 0.6
31 lambda: 1.05
32 - model: TareksLab/M-MERGE1
33 parameters:
34 weight:
35 - filter: self_attn
36 value: [0.2, 0.2, 0.1]
37 - filter: mlp
38 value: [0.1, 0.2, 0.2]
39 - value: 0.2
40 density: 0.6
41 lambda: 1
42 - model: TareksLab/M-BASE-SCE
43 parameters:
44 weight:
45 - filter: self_attn
46 value: [0.1, 0.3, 0.3]
47 - filter: mlp
48 value: [0.1, 0.2, 0.4]
49 - value: 0.2
50 density: 0.55
51 lambda: 1
52base_model: TareksLab/M-BASE-SCE
53merge_method: dare_ties
54parameters:
55 normalize: false
56 pad_to_multiple_of: 4
57tokenizer:
58 source: TareksLab/M-TOKENIZER-SCE
59chat_template: llama3
60dtype: bfloat16