NekoMikoReimu
commited on
Commit
•
35912f8
1
Parent(s):
c704d15
Update README.md
Browse files
README.md
CHANGED
@@ -1,36 +1,43 @@
|
|
1 |
---
|
2 |
license: apache-2.0
|
3 |
base_model: cyberagent/calm2-7b-chat
|
4 |
-
tags:
|
5 |
-
- generated_from_trainer
|
6 |
model-index:
|
7 |
- name: out
|
8 |
results: []
|
|
|
|
|
9 |
---
|
10 |
|
11 |
-
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
12 |
-
should probably proofread and complete it, then remove this comment. -->
|
13 |
|
14 |
-
[<img src="https://raw.githubusercontent.com/OpenAccess-AI-Collective/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/OpenAccess-AI-Collective/axolotl)
|
15 |
-
# out
|
16 |
-
|
17 |
-
This model is a fine-tuned version of [cyberagent/calm2-7b-chat](https://huggingface.co/cyberagent/calm2-7b-chat) on the None dataset.
|
18 |
-
It achieves the following results on the evaluation set:
|
19 |
-
- Loss: 3.2941
|
20 |
|
21 |
## Model description
|
22 |
|
23 |
-
|
|
|
|
|
|
|
|
|
24 |
|
25 |
## Intended uses & limitations
|
26 |
|
27 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
28 |
|
29 |
## Training and evaluation data
|
30 |
|
31 |
-
|
|
|
32 |
|
33 |
-
## Training procedure
|
34 |
|
35 |
### Training hyperparameters
|
36 |
|
@@ -49,35 +56,10 @@ The following hyperparameters were used during training:
|
|
49 |
- lr_scheduler_warmup_steps: 100
|
50 |
- num_epochs: 1
|
51 |
|
52 |
-
### Training results
|
53 |
-
|
54 |
-
| Training Loss | Epoch | Step | Validation Loss |
|
55 |
-
|:-------------:|:-----:|:----:|:---------------:|
|
56 |
-
| 3.5899 | 0.0 | 1 | 4.1781 |
|
57 |
-
| 3.4689 | 0.05 | 352 | 3.5988 |
|
58 |
-
| 3.519 | 0.1 | 704 | 3.6209 |
|
59 |
-
| 3.4168 | 0.15 | 1056 | 3.6161 |
|
60 |
-
| 3.4906 | 0.2 | 1408 | 3.6054 |
|
61 |
-
| 3.4853 | 0.25 | 1760 | 3.5950 |
|
62 |
-
| 3.475 | 0.3 | 2112 | 3.5806 |
|
63 |
-
| 3.2476 | 0.35 | 2464 | 3.5675 |
|
64 |
-
| 3.2339 | 0.4 | 2816 | 3.5398 |
|
65 |
-
| 3.3711 | 0.45 | 3168 | 3.5233 |
|
66 |
-
| 3.4195 | 0.5 | 3520 | 3.5006 |
|
67 |
-
| 3.1105 | 0.55 | 3872 | 3.4804 |
|
68 |
-
| 3.2477 | 0.6 | 4224 | 3.4546 |
|
69 |
-
| 3.2753 | 0.65 | 4576 | 3.4311 |
|
70 |
-
| 3.1867 | 0.7 | 4928 | 3.4054 |
|
71 |
-
| 3.2009 | 0.75 | 5280 | 3.3783 |
|
72 |
-
| 3.1037 | 0.8 | 5632 | 3.3509 |
|
73 |
-
| 3.0394 | 0.85 | 5984 | 3.3314 |
|
74 |
-
| 3.0759 | 0.9 | 6336 | 3.3095 |
|
75 |
-
| 3.1989 | 0.95 | 6688 | 3.2941 |
|
76 |
-
|
77 |
|
78 |
### Framework versions
|
79 |
|
80 |
- Transformers 4.34.1
|
81 |
- Pytorch 2.0.1+cu118
|
82 |
- Datasets 2.14.6
|
83 |
-
- Tokenizers 0.14.1
|
|
|
1 |
---
|
2 |
license: apache-2.0
|
3 |
base_model: cyberagent/calm2-7b-chat
|
|
|
|
|
4 |
model-index:
|
5 |
- name: out
|
6 |
results: []
|
7 |
+
language:
|
8 |
+
- ja
|
9 |
---
|
10 |
|
|
|
|
|
11 |
|
|
|
|
|
|
|
|
|
|
|
|
|
12 |
|
13 |
## Model description
|
14 |
|
15 |
+
Cyberagent様の[cyberagent/calm2-7b-chat](https://huggingface.co/cyberagent/calm2-7b-chat)を追加学習した、作家さん用アシスタントAIのアルファー版です。
|
16 |
+
文章を入れると、続きを書いてくれます。
|
17 |
+
まだまだトレーニングの途中ですけど、概念実証のためHFにアップロードしました。
|
18 |
+
これからどんどんクオリティーを向上させていく予定です。
|
19 |
+
|
20 |
|
21 |
## Intended uses & limitations
|
22 |
|
23 |
+
まだアルファー版ですけど、こちらの実験では十分アシスタントの役割を果たすことができることがわかりましたので、興味がある人はぜひお試しください。
|
24 |
+
[TextGen-WebUI](https://github.com/oobabooga/text-generation-webui)で使えます。
|
25 |
+
プロンプトは以下をお使いください。
|
26 |
+
```
|
27 |
+
あなたは誠実で優秀な日本人の書き手です。USERが書いた文章の続きを書いてください。
|
28 |
+
USER: (続きを書いてほしい文章)
|
29 |
+
ASSISTANT:
|
30 |
+
```
|
31 |
+
|
32 |
+
一例を添付します。
|
33 |
+
|
34 |
+
<img src="https://i.ibb.co/LnqfdJm/opennovel-alpha-screenshot.png" alt="opennovel-alpha-screenshot" border="0"></a>
|
35 |
|
36 |
## Training and evaluation data
|
37 |
|
38 |
+
約150Mトークンの小説テキストをShareGPT形式に変換して、追加学習しました。
|
39 |
+
|
40 |
|
|
|
41 |
|
42 |
### Training hyperparameters
|
43 |
|
|
|
56 |
- lr_scheduler_warmup_steps: 100
|
57 |
- num_epochs: 1
|
58 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
59 |
|
60 |
### Framework versions
|
61 |
|
62 |
- Transformers 4.34.1
|
63 |
- Pytorch 2.0.1+cu118
|
64 |
- Datasets 2.14.6
|
65 |
+
- Tokenizers 0.14.1
|