Hugging Face's logo

Nexesenex
/
Llama_3.x_70b_Tess_Dolphin_128K_v1.2

Text Generation
Transformers
Safetensors
llama
mergekit
Merge
conversational
text-generation-inference
Model card Files
xet
Community
Llama_3.x_70b_Tess_Dolphin_128K_v1.2
File size: 4,471 Bytes 
913213d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
 
merge_method: della_linear
base_model: migtissera/Tess-3-Llama-3.1-70B
models:
  - model: cognitivecomputations/dolphin-2.9.1-llama-3-70b
    parameters:
      weight:
        - filter: q_proj
          value: [0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0]
        - filter: k_proj
          value: [0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0]
        - filter: v_proj
          value: [0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0]
        - filter: o_proj
          value: [0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0]
        - filter: input_layernorm
          value: [0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0]
        - filter: up_proj
          value: [0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0]
        - filter: gate_proj
          value: [0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0]
        - filter: down_proj
          value: [0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0]
        - filter: post_attention_layernorm
          value: [0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0]
        - value: 0
      density: 0.5
      epsilon: 0.1
      lambda: 1.0
  - model: migtissera/Tess-3-Llama-3.1-70B
    parameters:
        weight: 1.0
        density:
          - filter: q_proj
            value: [1, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 1]
          - filter: k_proj
            value: [1, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 1]
          - filter: v_proj
            value: [1, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 1]
          - filter: o_proj
            value: [1, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 1]
          - filter: input_layernorm
            value: [1, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 1]
          - filter: up_proj
            value: [1, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 1]
          - filter: gate_proj
            value: [1, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 1]
          - filter: down_proj
            value: [1, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 1]
          - filter: post_attention_layernorm
            value: [1, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 1]
          - value: 0.5
        epsilon:
          - filter: q_proj
            value: [0, 0.01, 0.02, 0.03, 0.04, 0.05, 0.06, 0.07, 0.08, 0.09, 0.1, 0,09, 0.08, 0,07, 0.06, 0.05, 0.04, 0.03, 0.02, 0.01, 0]
          - filter: k_proj
            value: [0, 0.01, 0.02, 0.03, 0.04, 0.05, 0.06, 0.07, 0.08, 0.09, 0.1, 0,09, 0.08, 0,07, 0.06, 0.05, 0.04, 0.03, 0.02, 0.01, 0]
          - filter: v_proj
            value: [0, 0.01, 0.02, 0.03, 0.04, 0.05, 0.06, 0.07, 0.08, 0.09, 0.1, 0,09, 0.08, 0,07, 0.06, 0.05, 0.04, 0.03, 0.02, 0.01, 0]
          - filter: o_proj
            value: [0, 0.01, 0.02, 0.03, 0.04, 0.05, 0.06, 0.07, 0.08, 0.09, 0.1, 0,09, 0.08, 0,07, 0.06, 0.05, 0.04, 0.03, 0.02, 0.01, 0]
          - filter: input_layernorm
            value: [0, 0.01, 0.02, 0.03, 0.04, 0.05, 0.06, 0.07, 0.08, 0.09, 0.1, 0,09, 0.08, 0,07, 0.06, 0.05, 0.04, 0.03, 0.02, 0.01, 0]
          - filter: up_proj
            value: [0, 0.01, 0.02, 0.03, 0.04, 0.05, 0.06, 0.07, 0.08, 0.09, 0.1, 0,09, 0.08, 0,07, 0.06, 0.05, 0.04, 0.03, 0.02, 0.01, 0]
          - filter: gate_proj
            value: [0, 0.01, 0.02, 0.03, 0.04, 0.05, 0.06, 0.07, 0.08, 0.09, 0.1, 0,09, 0.08, 0,07, 0.06, 0.05, 0.04, 0.03, 0.02, 0.01, 0]
          - filter: down_proj
            value: [0, 0.01, 0.02, 0.03, 0.04, 0.05, 0.06, 0.07, 0.08, 0.09, 0.1, 0,09, 0.08, 0,07, 0.06, 0.05, 0.04, 0.03, 0.02, 0.01, 0]
          - filter: post_attention_layernorm
            value: [0, 0.01, 0.02, 0.03, 0.04, 0.05, 0.06, 0.07, 0.08, 0.09, 0.1, 0,09, 0.08, 0,07, 0.06, 0.05, 0.04, 0.03, 0.02, 0.01, 0]
          - value: 0.1
        lambda: 1.0
dtype: bfloat16
out_dtype: bfloat16
parameters:
  int8_mask: true
  normalize: true
  rescale: true
  filter_wise: false
chat_template: auto
tokenizer:
  source: union