layout
Browse files- run.py +53 -26
- script/custom.js +2 -3
- style.css +0 -0
- target_dataset.txt +70 -0
- type/dataset_type.py +59 -3
run.py
CHANGED
@@ -1,7 +1,7 @@
|
|
1 |
import gradio as gr
|
2 |
from huggingface_hub import list_models
|
3 |
from datasets import load_dataset
|
4 |
-
from type.dataset_type import
|
5 |
|
6 |
js = """
|
7 |
function blockEnter(event) {
|
@@ -16,11 +16,17 @@ function blockEnter(event) {
|
|
16 |
}
|
17 |
"""
|
18 |
|
|
|
19 |
|
20 |
-
def
|
|
|
|
|
|
|
|
|
|
|
21 |
if profile is None:
|
22 |
return "プライベートデータセット取得のためにログインしてください。"
|
23 |
-
return f
|
24 |
|
25 |
|
26 |
def list_private_models(profile: gr.OAuthProfile | None, oauth_token: gr.OAuthToken | None) -> str:
|
@@ -59,39 +65,60 @@ def display_dataset(profile: gr.OAuthProfile | None, oauth_token: gr.OAuthToken
|
|
59 |
else:
|
60 |
return gr.update(visible=True, value="データセットのロードに失敗しました。"), None, None
|
61 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
62 |
|
63 |
-
with gr.Blocks() as demo:
|
64 |
gr.Markdown("# 自動生成データセット アノテーション for Tanuki 2Phase")
|
65 |
|
66 |
-
with gr.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
67 |
|
68 |
-
|
69 |
-
|
|
|
70 |
|
71 |
-
|
72 |
-
# m2 = gr.Markdown(scale=1)
|
73 |
-
data_load_btn = gr.Button("データセットを読み込む")
|
74 |
|
75 |
-
|
76 |
-
|
77 |
-
|
|
|
|
|
78 |
|
79 |
-
|
80 |
-
|
81 |
-
bad_btn = gr.Button("悪い")
|
82 |
|
83 |
-
dataset_display = gr.Markdown(visible=False)
|
84 |
-
question_text = gr.Textbox(label="質問: ", interactive=False)
|
85 |
-
answer_text = gr.Textbox(label="回答: 改行はShift+Enterです。", elem_id="answer", interactive=True)
|
86 |
|
87 |
-
|
88 |
-
display_dataset,
|
89 |
-
inputs=None,
|
90 |
-
outputs=[dataset_display, question_text, answer_text],
|
91 |
-
)
|
92 |
|
93 |
-
|
94 |
-
return f" {answer_text}"
|
95 |
|
96 |
|
97 |
if __name__ == "__main__":
|
|
|
1 |
import gradio as gr
|
2 |
from huggingface_hub import list_models
|
3 |
from datasets import load_dataset
|
4 |
+
from type.dataset_type import TanukiPhase2AnnotationDataset
|
5 |
|
6 |
js = """
|
7 |
function blockEnter(event) {
|
|
|
16 |
}
|
17 |
"""
|
18 |
|
19 |
+
target_dataset = TanukiPhase2AnnotationDataset()
|
20 |
|
21 |
+
def load_css():
|
22 |
+
with open("style.css", "r") as file:
|
23 |
+
css_content = file.read()
|
24 |
+
return css_content
|
25 |
+
|
26 |
+
def set_name(profile: gr.OAuthProfile | None) -> str:
|
27 |
if profile is None:
|
28 |
return "プライベートデータセット取得のためにログインしてください。"
|
29 |
+
return f'{profile.username}さん、よろしくお願いいたします。'
|
30 |
|
31 |
|
32 |
def list_private_models(profile: gr.OAuthProfile | None, oauth_token: gr.OAuthToken | None) -> str:
|
|
|
65 |
else:
|
66 |
return gr.update(visible=True, value="データセットのロードに失敗しました。"), None, None
|
67 |
|
68 |
+
def switch_theme(theme):
|
69 |
+
if theme == "Dark":
|
70 |
+
return gr.themes.Default()
|
71 |
+
else:
|
72 |
+
return gr.themes.Monochrome()
|
73 |
+
|
74 |
+
theme_ = gr.State("Light")
|
75 |
+
|
76 |
+
with gr.Blocks(theme=theme_, css=load_css()) as demo:
|
77 |
|
|
|
78 |
gr.Markdown("# 自動生成データセット アノテーション for Tanuki 2Phase")
|
79 |
|
80 |
+
with gr.Tab("アノテーション"):
|
81 |
+
|
82 |
+
def update_theme():
|
83 |
+
new_theme = "Dark" if theme_.value == "Light" else "Light"
|
84 |
+
theme_.value = new_theme
|
85 |
+
return switch_theme(new_theme)
|
86 |
+
|
87 |
+
with gr.Row(equal_height=True):
|
88 |
+
|
89 |
+
login_btn = gr.LoginButton(visible=True, scale=1)
|
90 |
+
logout_btn = gr.LogoutButton(visible=False, scale=1)
|
91 |
+
|
92 |
+
profile = gr.Markdown()
|
93 |
+
data_load_btn = gr.Button("データセットを読み込む")
|
94 |
+
theme_button = gr.Button("Switch Theme")
|
95 |
+
theme_button.click(fn=update_theme, outputs=None)
|
96 |
+
|
97 |
+
demo.load(set_name, inputs=None, outputs=profile)
|
98 |
+
demo.load(toggle_buttons, inputs=None, outputs=[login_btn, logout_btn])
|
99 |
+
|
100 |
+
dataset_display = gr.Markdown(visible=False)
|
101 |
+
question_text = gr.Textbox(label="質問: ", interactive=False)
|
102 |
|
103 |
+
with gr.Row(equal_height=True):
|
104 |
+
good_btn = gr.Button("良い")
|
105 |
+
bad_btn = gr.Button("悪い")
|
106 |
|
107 |
+
answer_text = gr.Textbox(label="回答: 改行はShift+Enterです。", elem_id="answer", interactive=True)
|
|
|
|
|
108 |
|
109 |
+
data_load_btn.click(
|
110 |
+
display_dataset,
|
111 |
+
inputs=None,
|
112 |
+
outputs=[dataset_display, question_text, answer_text],
|
113 |
+
)
|
114 |
|
115 |
+
def on_submit(answer_text):
|
116 |
+
return f" {answer_text}"
|
|
|
117 |
|
|
|
|
|
|
|
118 |
|
119 |
+
with gr.Tab("アノテ済みデータセット"):
|
|
|
|
|
|
|
|
|
120 |
|
121 |
+
gr.Textbox("データセットID", lines=1, placeholder="データセットIDを入力してください。")
|
|
|
122 |
|
123 |
|
124 |
if __name__ == "__main__":
|
script/custom.js
CHANGED
@@ -1,4 +1,3 @@
|
|
1 |
function prevent (event) {
|
2 |
-
alert('Hello, World!');
|
3 |
-
|
4 |
-
};
|
|
|
1 |
function prevent (event) {
|
2 |
+
// alert('Hello, World!');
|
3 |
+
};
|
|
style.css
ADDED
File without changes
|
target_dataset.txt
ADDED
@@ -0,0 +1,70 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
## hatakeyama-llm-team/WikiBookJa
|
2 |
+
fileSize 29.3MB
|
3 |
+
- index int64
|
4 |
+
- text string
|
5 |
+
|
6 |
+
## hatakeyama-llm-team/AutoGeneratedJapaneseQA
|
7 |
+
fileSize 33.1MB
|
8 |
+
- question string
|
9 |
+
- database string
|
10 |
+
- answer string
|
11 |
+
- score float64
|
12 |
+
|
13 |
+
## hatakeyama-llm-team/AutoGeneratedJapaneseQA-other
|
14 |
+
fileSize 12.9MB
|
15 |
+
- question string
|
16 |
+
- database string
|
17 |
+
- answer string
|
18 |
+
- score float64
|
19 |
+
|
20 |
+
## kanhatakeyama/AutoWikiQA
|
21 |
+
fileSize 232MB
|
22 |
+
- question string
|
23 |
+
- inst_question string
|
24 |
+
- inst_answer_0 string
|
25 |
+
- text string
|
26 |
+
- database string
|
27 |
+
- answer string
|
28 |
+
- answer_0 string
|
29 |
+
- answer_1 string
|
30 |
+
|
31 |
+
## kanhatakeyama/ChatbotArenaJaMixtral8x22b
|
32 |
+
fileSize 13.9MB
|
33 |
+
- question string
|
34 |
+
- inst_question string
|
35 |
+
- inst_answer_0 string
|
36 |
+
- text string
|
37 |
+
- database string
|
38 |
+
- answer string
|
39 |
+
|
40 |
+
# kanhatakeyama/OrcaJaMixtral8x22b
|
41 |
+
fileSize 1.24GB
|
42 |
+
- question string
|
43 |
+
- inst_question string
|
44 |
+
- inst_answer_0 string
|
45 |
+
- text string
|
46 |
+
- database string
|
47 |
+
- answer string
|
48 |
+
|
49 |
+
# kanhatakeyama/AutoMultiTurnByMixtral8x22b
|
50 |
+
fileSize 270MB
|
51 |
+
- q1 string
|
52 |
+
- a1 string
|
53 |
+
- q2 string
|
54 |
+
- a2 string
|
55 |
+
- database string
|
56 |
+
|
57 |
+
# kanhatakeyama/LogicalDatasetsByMixtral8x22b
|
58 |
+
fileSize 116MB
|
59 |
+
- question string
|
60 |
+
- answer string
|
61 |
+
|
62 |
+
## output_dataset
|
63 |
+
- id int64
|
64 |
+
- dataset string
|
65 |
+
- dataset_id int64
|
66 |
+
- who string
|
67 |
+
- good bool
|
68 |
+
- bad bool
|
69 |
+
- is_proofreading string
|
70 |
+
- t_proofreading string
|
type/dataset_type.py
CHANGED
@@ -1,10 +1,66 @@
|
|
1 |
# データセットのファイル構造に合わせて型定義
|
2 |
-
|
|
|
3 |
index: int
|
4 |
text: str
|
5 |
|
6 |
-
class
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
7 |
question: str
|
|
|
|
|
|
|
|
|
8 |
database: str
|
9 |
answer: str
|
10 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
# データセットのファイル構造に合わせて型定義
|
2 |
+
|
3 |
+
class WikiBookJa:
|
4 |
index: int
|
5 |
text: str
|
6 |
|
7 |
+
class AutoGeneratedJapaneseQA:
|
8 |
+
question: str
|
9 |
+
database: str
|
10 |
+
answer: str
|
11 |
+
score: float
|
12 |
+
|
13 |
+
class AutoGeneratedJapaneseQA_other:
|
14 |
+
question: str
|
15 |
+
database: str
|
16 |
+
answer: str
|
17 |
+
score: float
|
18 |
+
|
19 |
+
class AutoWikiQA:
|
20 |
question: str
|
21 |
+
answer: str
|
22 |
+
inst_question: str
|
23 |
+
inst_answer_0: str
|
24 |
+
text : str
|
25 |
database: str
|
26 |
answer: str
|
27 |
+
answer_0: str
|
28 |
+
answer_1: str
|
29 |
+
|
30 |
+
class ChatbotArenaJaMixtral8x22b:
|
31 |
+
question: str
|
32 |
+
inst_question: str
|
33 |
+
inst_answer_0: str
|
34 |
+
text: str
|
35 |
+
database: str
|
36 |
+
answer: str
|
37 |
+
|
38 |
+
class OrcaJaMixtral8x22b:
|
39 |
+
question: str
|
40 |
+
inst_question: str
|
41 |
+
inst_answer_0: str
|
42 |
+
text: str
|
43 |
+
database: str
|
44 |
+
answer: str
|
45 |
+
|
46 |
+
class AutoMultiTurnByMixtral8x22b:
|
47 |
+
q1: str
|
48 |
+
a1: str
|
49 |
+
q2: str
|
50 |
+
a2: str
|
51 |
+
database: str
|
52 |
+
|
53 |
+
class LogicalDatasetsByMixtral8x22b:
|
54 |
+
question: str
|
55 |
+
answer: str
|
56 |
+
|
57 |
+
|
58 |
+
class TanukiPhase2AnnotationDataset:
|
59 |
+
id: int
|
60 |
+
dataset: str
|
61 |
+
database_id: int
|
62 |
+
who: str
|
63 |
+
good: bool
|
64 |
+
bad: bool
|
65 |
+
is_proofreading: bool
|
66 |
+
t_proofreading: str
|