Views
No views yet
1dtype: bfloat16
2merge_method: passthrough
3slices:
4- sources:
5 - layer_range: [0, 10]
6 model: Qwen/Qwen2.5-1.5B-Instruct
7- sources:
8 - layer_range: [10, 11]
9 model: Qwen/Qwen2.5-1.5B-Instruct
10 parameters:
11 scale:
12 - filter: o_proj
13 value: 0.1
14 - filter: down_proj
15 value: 0.1
16 - value: 1.0
17- sources:
18 - layer_range: [10, 11]
19 model: Qwen/Qwen2.5-1.5B-Instruct
20- sources:
21 - layer_range: [10, 11]
22 model: Qwen/Qwen2.5-1.5B-Instruct
23 parameters:
24 scale:
25 - filter: o_proj
26 value: 0.1
27 - filter: down_proj
28 value: 0.1
29 - value: 1.0
30- sources:
31 - layer_range: [11, 12]
32 model: Qwen/Qwen2.5-1.5B-Instruct
33 parameters:
34 scale:
35 - filter: o_proj
36 value: 0.1
37 - filter: down_proj
38 value: 0.1
39 - value: 1.0
40- sources:
41 - layer_range: [11, 12]
42 model: Qwen/Qwen2.5-1.5B-Instruct
43- sources:
44 - layer_range: [11, 12]
45 model: Qwen/Qwen2.5-1.5B-Instruct
46 parameters:
47 scale:
48 - filter: o_proj
49 value: 0.1
50 - filter: down_proj
51 value: 0.1
52 - value: 1.0
53- sources:
54 - layer_range: [12, 13]
55 model: Qwen/Qwen2.5-1.5B-Instruct
56 parameters:
57 scale:
58 - filter: o_proj
59 value: 0.1
60 - filter: down_proj
61 value: 0.1
62 - value: 1.0
63- sources:
64 - layer_range: [12, 13]
65 model: Qwen/Qwen2.5-1.5B-Instruct
66- sources:
67 - layer_range: [12, 13]
68 model: Qwen/Qwen2.5-1.5B-Instruct
69 parameters:
70 scale:
71 - filter: o_proj
72 value: 0.1
73 - filter: down_proj
74 value: 0.1
75 - value: 1.0
76- sources:
77 - layer_range: [13, 14]
78 model: Qwen/Qwen2.5-1.5B-Instruct
79 parameters:
80 scale:
81 - filter: o_proj
82 value: 0.1
83 - filter: down_proj
84 value: 0.1
85 - value: 1.0
86- sources:
87 - layer_range: [13, 14]
88 model: Qwen/Qwen2.5-1.5B-Instruct
89- sources:
90 - layer_range: [13, 14]
91 model: Qwen/Qwen2.5-1.5B-Instruct
92 parameters:
93 scale:
94 - filter: o_proj
95 value: 0.1
96 - filter: down_proj
97 value: 0.1
98 - value: 1.0
99- sources:
100 - layer_range: [14, 15]
101 model: Qwen/Qwen2.5-1.5B-Instruct
102 parameters:
103 scale:
104 - filter: o_proj
105 value: 0.1
106 - filter: down_proj
107 value: 0.1
108 - value: 1.0
109- sources:
110 - layer_range: [14, 15]
111 model: Qwen/Qwen2.5-1.5B-Instruct
112- sources:
113 - layer_range: [14, 15]
114 model: Qwen/Qwen2.5-1.5B-Instruct
115 parameters:
116 scale:
117 - filter: o_proj
118 value: 0.1
119 - filter: down_proj
120 value: 0.1
121 - value: 1.0
122- sources:
123 - layer_range: [15, 16]
124 model: Qwen/Qwen2.5-1.5B-Instruct
125 parameters:
126 scale:
127 - filter: o_proj
128 value: 0.1
129 - filter: down_proj
130 value: 0.1
131 - value: 1.0
132- sources:
133 - layer_range: [15, 16]
134 model: Qwen/Qwen2.5-1.5B-Instruct
135- sources:
136 - layer_range: [15, 16]
137 model: Qwen/Qwen2.5-1.5B-Instruct
138 parameters:
139 scale:
140 - filter: o_proj
141 value: 0.1
142 - filter: down_proj
143 value: 0.1
144 - value: 1.0
145- sources:
146 - layer_range: [16, 17]
147 model: Qwen/Qwen2.5-1.5B-Instruct
148 parameters:
149 scale:
150 - filter: o_proj
151 value: 0.1
152 - filter: down_proj
153 value: 0.1
154 - value: 1.0
155- sources:
156 - layer_range: [16, 17]
157 model: Qwen/Qwen2.5-1.5B-Instruct
158- sources:
159 - layer_range: [16, 17]
160 model: Qwen/Qwen2.5-1.5B-Instruct
161 parameters:
162 scale:
163 - filter: o_proj
164 value: 0.1
165 - filter: down_proj
166 value: 0.1
167 - value: 1.0
168- sources:
169 - layer_range: [17, 18]
170 model: Qwen/Qwen2.5-1.5B-Instruct
171 parameters:
172 scale:
173 - filter: o_proj
174 value: 0.1
175 - filter: down_proj
176 value: 0.1
177 - value: 1.0
178- sources:
179 - layer_range: [17, 18]
180 model: Qwen/Qwen2.5-1.5B-Instruct
181- sources:
182 - layer_range: [18, 28]
183 model: Qwen/Qwen2.5-1.5B-Instruct