Update README.md
Browse files
README.md
CHANGED
@@ -3,6 +3,8 @@ license: apache-2.0
|
|
3 |
language:
|
4 |
- ja
|
5 |
pipeline_tag: text-generation
|
|
|
|
|
6 |
---
|
7 |
# Japanese-Starling-ChatV-7B-GGUF
|
8 |
|
@@ -11,7 +13,7 @@ GGUF conversion of "<a href="https://huggingface.co/TFMC/Japanese-Starling-ChatV
|
|
11 |
"Japanese-Starling-ChatV-7B" is a Japanese chat model built on top of "<a href="https://huggingface.co/NTQAI/chatntq-ja-7b-v1.0">chatntq-ja-7b-v1.0</a>", originally based on Mistral-7B-v0.1.
|
12 |
I applied the chat vector acquired by subtracting the weights of Mistral-7B-v0.1 from the weights of "<a href="https://huggingface.co/Nexusflow/Starling-LM-7B-beta">Starling-LM-7B-beta</a>" to this model.
|
13 |
|
14 |
-
このモデルは
|
15 |
|
16 |
### Performance
|
17 |
|
@@ -25,12 +27,12 @@ I applied the chat vector acquired by subtracting the weights of Mistral-7B-v0.1
|
|
25 |
<th><a href="https://huggingface.co/mmnga/ELYZA-japanese-Llama-2-7b-instruct-gguf">ELYZA-japanese-Llama-2-7b-instruct-gguf</a></th>
|
26 |
</tr>
|
27 |
<tr>
|
28 |
-
<td>
|
29 |
-
<td
|
30 |
-
<td>Mistral
|
31 |
-
<td>Mistral
|
32 |
-
<td>Mistral
|
33 |
-
<td>Llama-2
|
34 |
</tr>
|
35 |
<tr>
|
36 |
<td>ELYZAtasks100<br>average score</td>
|
@@ -49,4 +51,4 @@ Scores on "<a href="https://huggingface.co/datasets/elyza/ELYZA-tasks-100">ELYZA
|
|
49 |
### Prompt Template
|
50 |
|
51 |
- Llama-2-Chat
|
52 |
-
<pre><code>[INST] <<SYS>>\nあなたは役に立つアシスタントです。\n<</SYS>>\n\n{prompt} [/INST]</code></pre>
|
|
|
3 |
language:
|
4 |
- ja
|
5 |
pipeline_tag: text-generation
|
6 |
+
tags:
|
7 |
+
- Mistral
|
8 |
---
|
9 |
# Japanese-Starling-ChatV-7B-GGUF
|
10 |
|
|
|
13 |
"Japanese-Starling-ChatV-7B" is a Japanese chat model built on top of "<a href="https://huggingface.co/NTQAI/chatntq-ja-7b-v1.0">chatntq-ja-7b-v1.0</a>", originally based on Mistral-7B-v0.1.
|
14 |
I applied the chat vector acquired by subtracting the weights of Mistral-7B-v0.1 from the weights of "<a href="https://huggingface.co/Nexusflow/Starling-LM-7B-beta">Starling-LM-7B-beta</a>" to this model.
|
15 |
|
16 |
+
このモデルはchatntq-ja-7b-v1.0をベースにした7Bパラメータの日本語チャットモデルです。高性能の英語モデルであるStarling-LM-7B-betaの重みからMistral-7B-v0.1の重みを差し引くことで得た"chat vector"を適用しています。
|
17 |
|
18 |
### Performance
|
19 |
|
|
|
27 |
<th><a href="https://huggingface.co/mmnga/ELYZA-japanese-Llama-2-7b-instruct-gguf">ELYZA-japanese-Llama-2-7b-instruct-gguf</a></th>
|
28 |
</tr>
|
29 |
<tr>
|
30 |
+
<td>Parameters</td>
|
31 |
+
<td>35B</td>
|
32 |
+
<td>7B(Mistral)</td>
|
33 |
+
<td>7B(Mistral)</td>
|
34 |
+
<td>7B(Mistral)</td>
|
35 |
+
<td>7B(Llama-2)</td>
|
36 |
</tr>
|
37 |
<tr>
|
38 |
<td>ELYZAtasks100<br>average score</td>
|
|
|
51 |
### Prompt Template
|
52 |
|
53 |
- Llama-2-Chat
|
54 |
+
<pre><code>[INST] <<SYS>>\nあなたは役に立つアシスタントです。\n<</SYS>>\n\n{prompt} [/INST]</code></pre>
|