Upload README.md
Browse files
README.md
ADDED
@@ -0,0 +1,173 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
base_model: []
|
3 |
+
library_name: transformers
|
4 |
+
tags:
|
5 |
+
- mergekit
|
6 |
+
- merge
|
7 |
+
|
8 |
+
---
|
9 |
+
# itr012_score3.93
|
10 |
+
|
11 |
+
This is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit).
|
12 |
+
|
13 |
+
## Merge Details
|
14 |
+
### Merge Method
|
15 |
+
|
16 |
+
This model was merged using the [linear](https://arxiv.org/abs/2203.05482) merge method using /home/umiyuki/automerge/evol_merge_storage/input_models/Japanese-Chat-Umievo-itr004-7b_579282327 as a base.
|
17 |
+
|
18 |
+
### Models Merged
|
19 |
+
|
20 |
+
The following models were included in the merge:
|
21 |
+
* /home/umiyuki/automerge/evol_merge_storage/input_models/Japanese-Starling-ChatV-7B_1737576410
|
22 |
+
* /home/umiyuki/automerge/evol_merge_storage/input_models/Ninja-v1-RP-expressive-v2_4102792561
|
23 |
+
* /home/umiyuki/automerge/evol_merge_storage/input_models/Vecteus-v1_4179808746
|
24 |
+
|
25 |
+
### Configuration
|
26 |
+
|
27 |
+
The following YAML configuration was used to produce this model:
|
28 |
+
|
29 |
+
```yaml
|
30 |
+
base_model: /home/umiyuki/automerge/evol_merge_storage/input_models/Japanese-Chat-Umievo-itr004-7b_579282327
|
31 |
+
dtype: bfloat16
|
32 |
+
merge_method: linear
|
33 |
+
parameters:
|
34 |
+
int8_mask: 1.0
|
35 |
+
normalize: 1.0
|
36 |
+
slices:
|
37 |
+
- sources:
|
38 |
+
- layer_range: [0, 4]
|
39 |
+
model: /home/umiyuki/automerge/evol_merge_storage/input_models/Japanese-Chat-Umievo-itr004-7b_579282327
|
40 |
+
parameters:
|
41 |
+
weight: 0.34953096474223655
|
42 |
+
- layer_range: [0, 4]
|
43 |
+
model: /home/umiyuki/automerge/evol_merge_storage/input_models/Vecteus-v1_4179808746
|
44 |
+
parameters:
|
45 |
+
weight: 0.4701212555597746
|
46 |
+
- layer_range: [0, 4]
|
47 |
+
model: /home/umiyuki/automerge/evol_merge_storage/input_models/Japanese-Starling-ChatV-7B_1737576410
|
48 |
+
parameters:
|
49 |
+
weight: 0.08162258723819021
|
50 |
+
- layer_range: [0, 4]
|
51 |
+
model: /home/umiyuki/automerge/evol_merge_storage/input_models/Ninja-v1-RP-expressive-v2_4102792561
|
52 |
+
parameters:
|
53 |
+
weight: 0.31015439852818116
|
54 |
+
- sources:
|
55 |
+
- layer_range: [4, 8]
|
56 |
+
model: /home/umiyuki/automerge/evol_merge_storage/input_models/Japanese-Chat-Umievo-itr004-7b_579282327
|
57 |
+
parameters:
|
58 |
+
weight: 0.11807412349683076
|
59 |
+
- layer_range: [4, 8]
|
60 |
+
model: /home/umiyuki/automerge/evol_merge_storage/input_models/Vecteus-v1_4179808746
|
61 |
+
parameters:
|
62 |
+
weight: -0.005684817244530085
|
63 |
+
- layer_range: [4, 8]
|
64 |
+
model: /home/umiyuki/automerge/evol_merge_storage/input_models/Japanese-Starling-ChatV-7B_1737576410
|
65 |
+
parameters:
|
66 |
+
weight: 0.2119283777941045
|
67 |
+
- layer_range: [4, 8]
|
68 |
+
model: /home/umiyuki/automerge/evol_merge_storage/input_models/Ninja-v1-RP-expressive-v2_4102792561
|
69 |
+
parameters:
|
70 |
+
weight: 1.1521124768396636
|
71 |
+
- sources:
|
72 |
+
- layer_range: [8, 12]
|
73 |
+
model: /home/umiyuki/automerge/evol_merge_storage/input_models/Japanese-Chat-Umievo-itr004-7b_579282327
|
74 |
+
parameters:
|
75 |
+
weight: 0.9244329405120573
|
76 |
+
- layer_range: [8, 12]
|
77 |
+
model: /home/umiyuki/automerge/evol_merge_storage/input_models/Vecteus-v1_4179808746
|
78 |
+
parameters:
|
79 |
+
weight: 0.7633842909616317
|
80 |
+
- layer_range: [8, 12]
|
81 |
+
model: /home/umiyuki/automerge/evol_merge_storage/input_models/Japanese-Starling-ChatV-7B_1737576410
|
82 |
+
parameters:
|
83 |
+
weight: 0.6952382990160072
|
84 |
+
- layer_range: [8, 12]
|
85 |
+
model: /home/umiyuki/automerge/evol_merge_storage/input_models/Ninja-v1-RP-expressive-v2_4102792561
|
86 |
+
parameters:
|
87 |
+
weight: 0.6873040403268571
|
88 |
+
- sources:
|
89 |
+
- layer_range: [12, 16]
|
90 |
+
model: /home/umiyuki/automerge/evol_merge_storage/input_models/Japanese-Chat-Umievo-itr004-7b_579282327
|
91 |
+
parameters:
|
92 |
+
weight: 0.4109625320908857
|
93 |
+
- layer_range: [12, 16]
|
94 |
+
model: /home/umiyuki/automerge/evol_merge_storage/input_models/Vecteus-v1_4179808746
|
95 |
+
parameters:
|
96 |
+
weight: 0.7090818691683626
|
97 |
+
- layer_range: [12, 16]
|
98 |
+
model: /home/umiyuki/automerge/evol_merge_storage/input_models/Japanese-Starling-ChatV-7B_1737576410
|
99 |
+
parameters:
|
100 |
+
weight: 0.42059423827890385
|
101 |
+
- layer_range: [12, 16]
|
102 |
+
model: /home/umiyuki/automerge/evol_merge_storage/input_models/Ninja-v1-RP-expressive-v2_4102792561
|
103 |
+
parameters:
|
104 |
+
weight: 0.5705186152354104
|
105 |
+
- sources:
|
106 |
+
- layer_range: [16, 20]
|
107 |
+
model: /home/umiyuki/automerge/evol_merge_storage/input_models/Japanese-Chat-Umievo-itr004-7b_579282327
|
108 |
+
parameters:
|
109 |
+
weight: 0.28507448659933315
|
110 |
+
- layer_range: [16, 20]
|
111 |
+
model: /home/umiyuki/automerge/evol_merge_storage/input_models/Vecteus-v1_4179808746
|
112 |
+
parameters:
|
113 |
+
weight: 0.4025223854083849
|
114 |
+
- layer_range: [16, 20]
|
115 |
+
model: /home/umiyuki/automerge/evol_merge_storage/input_models/Japanese-Starling-ChatV-7B_1737576410
|
116 |
+
parameters:
|
117 |
+
weight: 0.25885405316835886
|
118 |
+
- layer_range: [16, 20]
|
119 |
+
model: /home/umiyuki/automerge/evol_merge_storage/input_models/Ninja-v1-RP-expressive-v2_4102792561
|
120 |
+
parameters:
|
121 |
+
weight: 0.35540632690403373
|
122 |
+
- sources:
|
123 |
+
- layer_range: [20, 24]
|
124 |
+
model: /home/umiyuki/automerge/evol_merge_storage/input_models/Japanese-Chat-Umievo-itr004-7b_579282327
|
125 |
+
parameters:
|
126 |
+
weight: 0.018882795552694703
|
127 |
+
- layer_range: [20, 24]
|
128 |
+
model: /home/umiyuki/automerge/evol_merge_storage/input_models/Vecteus-v1_4179808746
|
129 |
+
parameters:
|
130 |
+
weight: 0.628847855051209
|
131 |
+
- layer_range: [20, 24]
|
132 |
+
model: /home/umiyuki/automerge/evol_merge_storage/input_models/Japanese-Starling-ChatV-7B_1737576410
|
133 |
+
parameters:
|
134 |
+
weight: 0.7038654876125734
|
135 |
+
- layer_range: [20, 24]
|
136 |
+
model: /home/umiyuki/automerge/evol_merge_storage/input_models/Ninja-v1-RP-expressive-v2_4102792561
|
137 |
+
parameters:
|
138 |
+
weight: 0.877501753107237
|
139 |
+
- sources:
|
140 |
+
- layer_range: [24, 28]
|
141 |
+
model: /home/umiyuki/automerge/evol_merge_storage/input_models/Japanese-Chat-Umievo-itr004-7b_579282327
|
142 |
+
parameters:
|
143 |
+
weight: 0.14008355431312197
|
144 |
+
- layer_range: [24, 28]
|
145 |
+
model: /home/umiyuki/automerge/evol_merge_storage/input_models/Vecteus-v1_4179808746
|
146 |
+
parameters:
|
147 |
+
weight: 1.0153826426873882
|
148 |
+
- layer_range: [24, 28]
|
149 |
+
model: /home/umiyuki/automerge/evol_merge_storage/input_models/Japanese-Starling-ChatV-7B_1737576410
|
150 |
+
parameters:
|
151 |
+
weight: 0.5586634927008272
|
152 |
+
- layer_range: [24, 28]
|
153 |
+
model: /home/umiyuki/automerge/evol_merge_storage/input_models/Ninja-v1-RP-expressive-v2_4102792561
|
154 |
+
parameters:
|
155 |
+
weight: 0.54455848971032
|
156 |
+
- sources:
|
157 |
+
- layer_range: [28, 32]
|
158 |
+
model: /home/umiyuki/automerge/evol_merge_storage/input_models/Japanese-Chat-Umievo-itr004-7b_579282327
|
159 |
+
parameters:
|
160 |
+
weight: 0.8188405381342685
|
161 |
+
- layer_range: [28, 32]
|
162 |
+
model: /home/umiyuki/automerge/evol_merge_storage/input_models/Vecteus-v1_4179808746
|
163 |
+
parameters:
|
164 |
+
weight: 0.5130358379308082
|
165 |
+
- layer_range: [28, 32]
|
166 |
+
model: /home/umiyuki/automerge/evol_merge_storage/input_models/Japanese-Starling-ChatV-7B_1737576410
|
167 |
+
parameters:
|
168 |
+
weight: 1.1132727871460124
|
169 |
+
- layer_range: [28, 32]
|
170 |
+
model: /home/umiyuki/automerge/evol_merge_storage/input_models/Ninja-v1-RP-expressive-v2_4102792561
|
171 |
+
parameters:
|
172 |
+
weight: 0.4471258297582539
|
173 |
+
```
|