Views
No views yet
1merge_method: passthrough
2slices:
3- sources:
4 - layer_range: [0, 19]
5 model: Qwen/Qwen2.5-7B-Instruct
6- sources:
7 - layer_range: [19, 20]
8 model: Qwen/Qwen2.5-7B-Instruct
9- sources:
10 - layer_range: [19, 20]
11 model: Qwen/Qwen2.5-7B-Instruct
12 parameters:
13 scale:
14 - filter: o_proj
15 value: 0.0
16 - filter: down_proj
17 value: 0.0
18 - value: 1.0
19- sources:
20 - layer_range: [19, 20]
21 model: Qwen/Qwen2.5-7B-Instruct
22 parameters:
23 scale:
24 - filter: o_proj
25 value: 0.0
26 - filter: down_proj
27 value: 0.0
28 - value: 1.0
29- sources:
30 - layer_range: [20, 21]
31 model: Qwen/Qwen2.5-7B-Instruct
32- sources:
33 - layer_range: [20, 21]
34 model: Qwen/Qwen2.5-7B-Instruct
35 parameters:
36 scale:
37 - filter: o_proj
38 value: 0.0
39 - filter: down_proj
40 value: 0.0
41 - value: 1.0
42- sources:
43 - layer_range: [20, 21]
44 model: Qwen/Qwen2.5-7B-Instruct
45 parameters:
46 scale:
47 - filter: o_proj
48 value: 0.0
49 - filter: down_proj
50 value: 0.0
51 - value: 1.0
52- sources:
53 - layer_range: [21, 22]
54 model: Qwen/Qwen2.5-7B-Instruct
55- sources:
56 - layer_range: [21, 22]
57 model: Qwen/Qwen2.5-7B-Instruct
58 parameters:
59 scale:
60 - filter: o_proj
61 value: 0.0
62 - filter: down_proj
63 value: 0.0
64 - value: 1.0
65- sources:
66 - layer_range: [21, 22]
67 model: Qwen/Qwen2.5-7B-Instruct
68 parameters:
69 scale:
70 - filter: o_proj
71 value: 0.0
72 - filter: down_proj
73 value: 0.0
74 - value: 1.0
75- sources:
76 - layer_range: [22, 23]
77 model: Qwen/Qwen2.5-7B-Instruct
78- sources:
79 - layer_range: [22, 23]
80 model: Qwen/Qwen2.5-7B-Instruct
81 parameters:
82 scale:
83 - filter: o_proj
84 value: 0.0
85 - filter: down_proj
86 value: 0.0
87 - value: 1.0
88- sources:
89 - layer_range: [22, 23]
90 model: Qwen/Qwen2.5-7B-Instruct
91 parameters:
92 scale:
93 - filter: o_proj
94 value: 0.0
95 - filter: down_proj
96 value: 0.0
97 - value: 1.0
98- sources:
99 - layer_range: [23, 24]
100 model: Qwen/Qwen2.5-7B-Instruct
101- sources:
102 - layer_range: [23, 24]
103 model: Qwen/Qwen2.5-7B-Instruct
104 parameters:
105 scale:
106 - filter: o_proj
107 value: 0.0
108 - filter: down_proj
109 value: 0.0
110 - value: 1.0
111- sources:
112 - layer_range: [23, 24]
113 model: Qwen/Qwen2.5-7B-Instruct
114 parameters:
115 scale:
116 - filter: o_proj
117 value: 0.0
118 - filter: down_proj
119 value: 0.0
120 - value: 1.0
121- sources:
122 - layer_range: [24, 25]
123 model: Qwen/Qwen2.5-7B-Instruct
124- sources:
125 - layer_range: [24, 25]
126 model: Qwen/Qwen2.5-7B-Instruct
127 parameters:
128 scale:
129 - filter: o_proj
130 value: 0.0
131 - filter: down_proj
132 value: 0.0
133 - value: 1.0
134- sources:
135 - layer_range: [24, 25]
136 model: Qwen/Qwen2.5-7B-Instruct
137 parameters:
138 scale:
139 - filter: o_proj
140 value: 0.0
141 - filter: down_proj
142 value: 0.0
143 - value: 1.0
144- sources:
145 - layer_range: [25, 26]
146 model: Qwen/Qwen2.5-7B-Instruct
147- sources:
148 - layer_range: [25, 26]
149 model: Qwen/Qwen2.5-7B-Instruct
150 parameters:
151 scale:
152 - filter: o_proj
153 value: 0.0
154 - filter: down_proj
155 value: 0.0
156 - value: 1.0
157- sources:
158 - layer_range: [25, 26]
159 model: Qwen/Qwen2.5-7B-Instruct
160 parameters:
161 scale:
162 - filter: o_proj
163 value: 0.0
164 - filter: down_proj
165 value: 0.0
166 - value: 1.0
167- sources:
168 - layer_range: [26, 27]
169 model: Qwen/Qwen2.5-7B-Instruct
170- sources:
171 - layer_range: [26, 27]
172 model: Qwen/Qwen2.5-7B-Instruct
173 parameters:
174 scale:
175 - filter: o_proj
176 value: 0.0
177 - filter: down_proj
178 value: 0.0
179 - value: 1.0
180- sources:
181 - layer_range: [26, 27]
182 model: Qwen/Qwen2.5-7B-Instruct
183 parameters:
184 scale:
185 - filter: o_proj
186 value: 0.0
187 - filter: down_proj
188 value: 0.0
189 - value: 1.0
190- sources:
191 - layer_range: [27, 28]
192 model: Qwen/Qwen2.5-7B-Instruct
193- sources:
194 - layer_range: [27, 28]
195 model: Qwen/Qwen2.5-7B-Instruct
196 parameters:
197 scale:
198 - filter: o_proj
199 value: 0.0
200 - filter: down_proj
201 value: 0.0
202 - value: 1.0
203- sources:
204 - layer_range: [27, 28]
205 model: Qwen/Qwen2.5-7B-Instruct
206 parameters:
207 scale:
208 - filter: o_proj
209 value: 0.0
210 - filter: down_proj
211 value: 0.0
212 - value: 1.0