Nohobby commited on
Commit
101a936
1 Parent(s): a0fa329

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +83 -29
README.md CHANGED
@@ -17,54 +17,108 @@ This is a merge of pre-trained language models created using [mergekit](https://
17
 
18
  This model was merged using the della_linear merge method using [ArliAI/Mistral-Nemo-12B-ArliAI-RPMax-v1.1](https://huggingface.co/ArliAI/Mistral-Nemo-12B-ArliAI-RPMax-v1.1) as a base.
19
 
20
- ### Models Merged
21
-
22
- The following models were included in the merge:
23
- * Insanity/chatml
24
- * Insanity/uncen
25
- * [v000000/NM-12B-Lyris-dev-3](https://huggingface.co/v000000/NM-12B-Lyris-dev-3)
26
- * Insanity/conv
27
-
28
  ### Configuration
29
 
30
  The following YAML configuration was used to produce this model:
31
 
32
  ```yaml
33
- base_model: ArliAI/Mistral-Nemo-12B-ArliAI-RPMax-v1.1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
34
  dtype: bfloat16
35
- merge_method: della_linear
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
36
  parameters:
37
  epsilon: 0.04
38
- int8_mask: 1.0
39
  lambda: 1.05
40
- normalize: 0.0
41
- rescale: 1.0
42
- slices:
43
- - sources:
44
- - layer_range: [0, 40]
45
- model: ArliAI/Mistral-Nemo-12B-ArliAI-RPMax-v1.1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
46
  parameters:
47
- density: [0.45, 0.55, 0.45, 0.55, 0.45]
48
  weight: [0.2, 0.3, 0.2, 0.3, 0.2]
49
- - layer_range: [0, 40]
50
- model: Insanity/chatml
51
  parameters:
52
- density: [0.6, 0.4, 0.5, 0.4, 0.6]
53
  weight: [0.01768, -0.01675, 0.01285, -0.01696, 0.01421]
54
- - layer_range: [0, 40]
55
- model: Insanity/uncen
56
  parameters:
57
  density: [0.6, 0.4, 0.5, 0.4, 0.6]
58
  weight: [0.01768, -0.01675, 0.01285, -0.01696, 0.01421]
59
- - layer_range: [0, 40]
60
- model: Insanity/conv
61
  parameters:
62
- density: [0.7]
63
  weight: [0.208, 0.139, 0.139, 0.139, 0.208]
64
- - layer_range: [0, 40]
65
- model: v000000/NM-12B-Lyris-dev-3
66
  parameters:
67
- density: [0.45, 0.55, 0.45, 0.55, 0.45]
68
  weight: [0.33]
 
 
 
 
 
 
 
 
 
 
69
  tokenizer_source: base
70
  ```
 
17
 
18
  This model was merged using the della_linear merge method using [ArliAI/Mistral-Nemo-12B-ArliAI-RPMax-v1.1](https://huggingface.co/ArliAI/Mistral-Nemo-12B-ArliAI-RPMax-v1.1) as a base.
19
 
 
 
 
 
 
 
 
 
20
  ### Configuration
21
 
22
  The following YAML configuration was used to produce this model:
23
 
24
  ```yaml
25
+ models:
26
+ - model: natong19/Mistral-Nemo-Instruct-2407-abliterated
27
+ - model: Fizzarolli/MN-12b-Sunrose
28
+ parameters:
29
+ density: 0.5
30
+ weight: [0.495, 0.165, 0.165, 0.495, 0.495, 0.165, 0.165, 0.495]
31
+ - model: nbeerbower/mistral-nemo-gutenberg-12B-v4
32
+ parameters:
33
+ density: [0.35, 0.65, 0.5, 0.65, 0.35]
34
+ weight: [-0.01891, 0.01554, -0.01325, 0.01791, -0.01458]
35
+ merge_method: dare_ties
36
+ base_model: natong19/Mistral-Nemo-Instruct-2407-abliterated
37
+ parameters:
38
+ normalize: false
39
+ int8_mask: true
40
  dtype: bfloat16
41
+ name: uncen
42
+
43
+ ---
44
+ models:
45
+ - model: unsloth/Mistral-Nemo-Instruct-2407
46
+ - model: NeverSleep/Lumimaid-v0.2-12B
47
+ parameters:
48
+ density: 0.5
49
+ weight: [0.139, 0.208, 0.139, 0.208, 0.139]
50
+ - model: nbeerbower/mistral-nemo-cc-12B
51
+ parameters:
52
+ density: [0.65, 0.35, 0.5, 0.35, 0.65]
53
+ weight: [0.01823, -0.01647, 0.01422, -0.01975, 0.01128]
54
+ - model: nbeerbower/mistral-nemo-bophades-12B
55
+ parameters:
56
+ density: [0.35, 0.65, 0.5, 0.65, 0.35]
57
+ weight: [-0.01891, 0.01554, -0.01325, 0.01791, -0.01458]
58
+ merge_method: della
59
+ base_model: unsloth/Mistral-Nemo-Instruct-2407
60
  parameters:
61
  epsilon: 0.04
 
62
  lambda: 1.05
63
+ normalize: false
64
+ int8_mask: true
65
+ dtype: bfloat16
66
+ name: conv
67
+ ---
68
+ models:
69
+ - model: unsloth/Mistral-Nemo-Base-2407
70
+ - model: elinas/Chronos-Gold-12B-1.0
71
+ parameters:
72
+ density: 0.9
73
+ gamma: 0.01
74
+ weight: [0.139, 0.208, 0.208, 0.139, 0.139]
75
+ - model: shuttleai/shuttle-2.5-mini
76
+ parameters:
77
+ density: 0.9
78
+ gamma: 0.01
79
+ weight: [0.208, 0.139, 0.139, 0.139, 0.208]
80
+ - model: Epiculous/Violet_Twilight-v0.2
81
+ parameters:
82
+ density: 0.9
83
+ gamma: 0.01
84
+ weight: [0.139, 0.139, 0.208, 0.208, 0.139]
85
+ merge_method: breadcrumbs_ties
86
+ base_model: unsloth/Mistral-Nemo-Base-2407
87
+ parameters:
88
+ normalize: false
89
+ int8_mask: true
90
+ dtype: bfloat16
91
+ name: chatml
92
+ ---
93
+ models:
94
+ - model: ArliAI/Mistral-Nemo-12B-ArliAI-RPMax-v1.1
95
  parameters:
 
96
  weight: [0.2, 0.3, 0.2, 0.3, 0.2]
97
+ density: [0.45, 0.55, 0.45, 0.55, 0.45]
98
+ - model: chatml
99
  parameters:
 
100
  weight: [0.01768, -0.01675, 0.01285, -0.01696, 0.01421]
101
+ density: [0.6, 0.4, 0.5, 0.4, 0.6]
102
+ - model: uncen
103
  parameters:
104
  density: [0.6, 0.4, 0.5, 0.4, 0.6]
105
  weight: [0.01768, -0.01675, 0.01285, -0.01696, 0.01421]
106
+ - model: conv
 
107
  parameters:
 
108
  weight: [0.208, 0.139, 0.139, 0.139, 0.208]
109
+ density: [0.7]
110
+ - model: v000000/NM-12B-Lyris-dev-3
111
  parameters:
 
112
  weight: [0.33]
113
+ density: [0.45, 0.55, 0.45, 0.55, 0.45]
114
+ merge_method: della_linear
115
+ base_model: ArliAI/Mistral-Nemo-12B-ArliAI-RPMax-v1.1
116
+ parameters:
117
+ epsilon: 0.04
118
+ lambda: 1.05
119
+ int8_mask: true
120
+ rescale: true
121
+ normalize: false
122
+ dtype: bfloat16
123
  tokenizer_source: base
124
  ```