aashish1904 commited on
Commit
f14e432
1 Parent(s): 277f01f

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +171 -0
README.md ADDED
@@ -0,0 +1,171 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+ ---
3
+
4
+ license: apache-2.0
5
+ tags:
6
+ - merge
7
+ - mergekit
8
+ - MaziyarPanahi/Calme-7B-Instruct-v0.1.1
9
+
10
+ ---
11
+
12
+ ![](https://cdn.discordapp.com/attachments/791342238541152306/1264099835221381251/image.png?ex=669ca436&is=669b52b6&hm=129f56187c31e1ed22cbd1bcdbc677a2baeea5090761d2f1a458c8b1ec7cca4b&)
13
+
14
+ # QuantFactory/Calme-Instruct-Extended-GGUF
15
+ This is quantized version of [arcee-ai/Calme-Instruct-Extended](https://huggingface.co/arcee-ai/Calme-Instruct-Extended) created using llama.cpp
16
+
17
+ # Original Model Card
18
+
19
+
20
+ # Calme-Instruct-Extended
21
+
22
+ Calme-Instruct-Extended is a merge of the following models using [mergekit](https://github.com/cg123/mergekit):
23
+ * [MaziyarPanahi/Calme-7B-Instruct-v0.1.1](https://huggingface.co/MaziyarPanahi/Calme-7B-Instruct-v0.1.1)
24
+
25
+
26
+ ## 🧩 Configuration
27
+
28
+ ```yaml
29
+ slices:
30
+ - sources:
31
+ - model: MaziyarPanahi/Calme-7B-Instruct-v0.1.1
32
+ layer_range:
33
+ - 0
34
+ - 4
35
+ - sources:
36
+ - model: MaziyarPanahi/Calme-7B-Instruct-v0.1.1
37
+ layer_range:
38
+ - 3
39
+ - 4
40
+ parameters:
41
+ scale:
42
+ - filter: o_proj
43
+ value: 0
44
+ - filter: down_proj
45
+ value: 0
46
+ - value: 1
47
+ - sources:
48
+ - model: MaziyarPanahi/Calme-7B-Instruct-v0.1.1
49
+ layer_range:
50
+ - 4
51
+ - 8
52
+ - sources:
53
+ - model: MaziyarPanahi/Calme-7B-Instruct-v0.1.1
54
+ layer_range:
55
+ - 7
56
+ - 8
57
+ parameters:
58
+ scale:
59
+ - filter: o_proj
60
+ value: 0
61
+ - filter: down_proj
62
+ value: 0
63
+ - value: 1
64
+ - sources:
65
+ - model: MaziyarPanahi/Calme-7B-Instruct-v0.1.1
66
+ layer_range:
67
+ - 8
68
+ - 12
69
+ - sources:
70
+ - model: MaziyarPanahi/Calme-7B-Instruct-v0.1.1
71
+ layer_range:
72
+ - 11
73
+ - 12
74
+ parameters:
75
+ scale:
76
+ - filter: o_proj
77
+ value: 0
78
+ - filter: down_proj
79
+ value: 0
80
+ - value: 1
81
+ - sources:
82
+ - model: MaziyarPanahi/Calme-7B-Instruct-v0.1.1
83
+ layer_range:
84
+ - 12
85
+ - 16
86
+ - sources:
87
+ - model: MaziyarPanahi/Calme-7B-Instruct-v0.1.1
88
+ layer_range:
89
+ - 15
90
+ - 16
91
+ parameters:
92
+ scale:
93
+ - filter: o_proj
94
+ value: 0
95
+ - filter: down_proj
96
+ value: 0
97
+ - value: 1
98
+ - sources:
99
+ - model: MaziyarPanahi/Calme-7B-Instruct-v0.1.1
100
+ layer_range:
101
+ - 16
102
+ - 20
103
+ - sources:
104
+ - model: MaziyarPanahi/Calme-7B-Instruct-v0.1.1
105
+ layer_range:
106
+ - 19
107
+ - 20
108
+ parameters:
109
+ scale:
110
+ - filter: o_proj
111
+ value: 0
112
+ - filter: down_proj
113
+ value: 0
114
+ - value: 1
115
+ - sources:
116
+ - model: MaziyarPanahi/Calme-7B-Instruct-v0.1.1
117
+ layer_range:
118
+ - 20
119
+ - 24
120
+ - sources:
121
+ - model: MaziyarPanahi/Calme-7B-Instruct-v0.1.1
122
+ layer_range:
123
+ - 23
124
+ - 24
125
+ parameters:
126
+ scale:
127
+ - filter: o_proj
128
+ value: 0
129
+ - filter: down_proj
130
+ value: 0
131
+ - value: 1
132
+ - sources:
133
+ - model: MaziyarPanahi/Calme-7B-Instruct-v0.1.1
134
+ layer_range:
135
+ - 24
136
+ - 28
137
+ - sources:
138
+ - model: MaziyarPanahi/Calme-7B-Instruct-v0.1.1
139
+ layer_range:
140
+ - 27
141
+ - 28
142
+ parameters:
143
+ scale:
144
+ - filter: o_proj
145
+ value: 0
146
+ - filter: down_proj
147
+ value: 0
148
+ - value: 1
149
+ - sources:
150
+ - model: MaziyarPanahi/Calme-7B-Instruct-v0.1.1
151
+ layer_range:
152
+ - 28
153
+ - 32
154
+ - sources:
155
+ - model: MaziyarPanahi/Calme-7B-Instruct-v0.1.1
156
+ layer_range:
157
+ - 31
158
+ - 32
159
+ parameters:
160
+ scale:
161
+ - filter: o_proj
162
+ value: 0
163
+ - filter: down_proj
164
+ value: 0
165
+ - value: 1
166
+ merge_method: passthrough
167
+ dtype: bfloat16
168
+
169
+
170
+
171
+ ```