Update README.md
Browse files
README.md
CHANGED
@@ -4,9 +4,24 @@ library_name: transformers
|
|
4 |
tags:
|
5 |
- mergekit
|
6 |
- merge
|
7 |
-
|
|
|
|
|
8 |
---
|
9 |
-
#
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
10 |
|
11 |
This is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit).
|
12 |
|
@@ -170,4 +185,4 @@ slices:
|
|
170 |
model: /home/umiyuki/automerge/evol_merge_storage/input_models/Ninja-v1-RP-expressive-v2_4102792561
|
171 |
parameters:
|
172 |
weight: 0.4471258297582539
|
173 |
-
```
|
|
|
4 |
tags:
|
5 |
- mergekit
|
6 |
- merge
|
7 |
+
license: apache-2.0
|
8 |
+
language:
|
9 |
+
- ja
|
10 |
---
|
11 |
+
# Umievo-itr012-Gleipnir-7B
|
12 |
+
|
13 |
+
このモデルは強力な4つの日本語モデルを進化的アルゴリズムで進化的マージしたものです。Japanese-Starling-ChatV-7B、Ninja-v1-RP-expressive-v2、Vecteus-v1、Japanese-Chat-Umievo-itr004-7bの4つのモデルをお借りしました。
|
14 |
+
マージに使用させていただいたモデル制作者のAratakoさん、Bakuさん、Local-Novel-LLM-projectのみなさまに感謝します。それから問題解決のきっかけをくれたHoly-foxさんに感謝します。
|
15 |
+
|
16 |
+
This model is an evolutionary merge of four powerful Japanese models with an evolutionary algorithm. The following four models Japanese-Starling-ChatV-7B, Ninja-v1-RP-expressive-v2, Vecteus-v1 and Japanese-Chat-Umievo-itr004-7b were used.
|
17 |
+
I would like to thank the model makers Aratako, Baku and Local-Novel-LLM-project for allowing me to use their models for the merge. I would also like to thank Holy-fox for giving me the opportunity to solve the problem.
|
18 |
+
|
19 |
+
ElyzaTasks100ベンチマークで平均点が3.91でした。(Llama3-70Bによる自動評価を3回行った平均点)
|
20 |
+
|
21 |
+
The average score was 3.91 on the ElyzaTasks100 benchmark. (Average score after 3 automatic evaluations by Llama3-70B)
|
22 |
+
|
23 |
+
![image/png](https://cdn-uploads.huggingface.co/production/uploads/630420b4eedc089484c853e8/FxUBzBUKpe_JSHSJufSv5.png)
|
24 |
+
|
25 |
|
26 |
This is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit).
|
27 |
|
|
|
185 |
model: /home/umiyuki/automerge/evol_merge_storage/input_models/Ninja-v1-RP-expressive-v2_4102792561
|
186 |
parameters:
|
187 |
weight: 0.4471258297582539
|
188 |
+
```
|