Hugging Face's logo

Spaces:
pmkhanh7890
/
handwriting-recognition
Sleeping

App Files Community
pmkhanh7890 commited on
Commit
6610027
·
1 Parent(s): 78a90b5

1st push

Browse files
Files changed (21) hide show
  1. .DS_Store +0 -0
  2. .gitignore +113 -0
  3. .pre-commit-config.yaml +53 -0
  4. .pre-commit-setting.toml +22 -0
  5. _config.yml +1 -0
  6. app.py +68 -0
  7. app_streamlit.py +104 -0
  8. data/in_1.png +0 -0
  9. data/sample_1_1.png +0 -0
  10. dict/japanese_charlist.txt +4441 -0
  11. models/handwritten-japanese-recognition-0001.bin +3 -0
  12. models/handwritten-japanese-recognition-0001.xml +2252 -0
  13. pyproject.toml +21 -0
  14. requirements.txt +4 -0
  15. src/__Init__.py +0 -0
  16. src/config.py +26 -0
  17. src/detect_mask_image.py +225 -0
  18. src/image_processing.py +60 -0
  19. src/model2onnx.py +41 -0
  20. src/search.py +94 -0
  21. src/train_mask_detector.py +192 -0
.DS_Store ADDED
Binary file (8.2 kB). View file
 
.gitignore ADDED
@@ -0,0 +1,113 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # virtual environment
2
+ .venv
3
+ common/cpp
4
+ common/cpp_gapi
5
+ sample_videos/*
6
+ models/vietocr/*
7
+ models/soundbar/*
8
+ results.json
9
+ output/*
10
+ models/soundbar_detection_yolov7/*
11
+ # Byte-compiled / optimized / DLL files
12
+ __pycache__/
13
+ *.py[cod]
14
+ *$py.class
15
+ # C extensions
16
+ *.so
17
+ # Distribution / packaging
18
+ .Python
19
+ build/
20
+ develop-eggs/
21
+ dist/
22
+ downloads/
23
+ eggs/
24
+ .eggs/
25
+ lib/
26
+ lib64/
27
+ parts/
28
+ sdist/
29
+ var/
30
+ wheels/
31
+ share/python-wheels/
32
+ *.egg-info/
33
+ .installed.cfg
34
+ *.egg
35
+ MANIFEST
36
+ # source/yolov7/*
37
+ # PyInstaller
38
+ # Usually these files are written by a python script from a template
39
+ # before PyInstaller builds the exe, so as to inject date/other infos into it.
40
+ *.manifest
41
+ *.spec
42
+ # Installer logs
43
+ pip-log.txt
44
+ pip-delete-this-directory.txt
45
+ # Unit test / coverage reports
46
+ htmlcov/
47
+ .tox/
48
+ .nox/
49
+ .coverage
50
+ .coverage.*
51
+ .cache
52
+ nosetests.xml
53
+ coverage.xml
54
+ *.cover
55
+ .hypothesis/
56
+ .pytest_cache/
57
+ # Translations
58
+ *.mo
59
+ *.pot
60
+ # Django stuff:
61
+ *.log
62
+ local_settings.py
63
+ db.sqlite3
64
+ # Flask stuff:
65
+ instance/
66
+ .webassets-cache
67
+ # Scrapy stuff:
68
+ .scrapy
69
+ # Sphinx documentation
70
+ docs/_build/
71
+ # PyBuilder
72
+ target/
73
+ # Jupyter Notebook
74
+ .ipynb_checkpoints
75
+ # IPython
76
+ profile_default/
77
+ ipython_config.py
78
+ # pyenv
79
+ .python-version
80
+ # celery beat schedule file
81
+ celerybeat-schedule
82
+ # SageMath parsed files
83
+ *.sage.py
84
+ # Environments
85
+ .env
86
+ .venv
87
+ env/
88
+ venv/
89
+ ENV/
90
+ env.bak/
91
+ venv.bak/
92
+ # Spyder project settings
93
+ .spyderproject
94
+ .spyproject
95
+ # Rope project settings
96
+ .ropeproject
97
+ # mkdocs documentation
98
+ /site
99
+ # mypy
100
+ .mypy_cache/
101
+ .dmypy.json
102
+ dmypy.json
103
+ # Pyre type checker
104
+ .pyre/
105
+ # custom
106
+ old/
107
+ notebook/
108
+ static/images/face
109
+ static/images/celeba
110
+ celeba_pics.gz
111
+ celeba_vecs.gz
112
+ face_pics.gz
113
+ face_vecs.gz
.pre-commit-config.yaml ADDED
@@ -0,0 +1,53 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # See https://pre-commit.com for more information
2
+ # See https://pre-commit.com/hooks.html for more hooks
3
+ repos:
4
+ - repo: https://github.com/pre-commit/pre-commit-hooks
5
+ rev: v4.2.0
6
+ hooks:
7
+ #- id: check-added-large-files
8
+ - id: fix-byte-order-marker
9
+ - id: check-case-conflict
10
+ - id: check-json
11
+ - id: check-yaml
12
+ args: ['--unsafe']
13
+ - id: detect-aws-credentials
14
+ args: [--allow-missing-credentials]
15
+ - id: detect-private-key
16
+ - id: end-of-file-fixer
17
+ - id: mixed-line-ending
18
+ - id: trailing-whitespace
19
+ - repo: https://github.com/asottile/add-trailing-comma
20
+ rev: v2.2.2
21
+ hooks:
22
+ - id: add-trailing-comma
23
+ - repo: https://github.com/pycqa/isort
24
+ rev: 5.10.1
25
+ hooks:
26
+ - id: isort
27
+ name: isort (python)
28
+ args: [--settings-path=pyproject.toml]
29
+ - id: isort
30
+ name: isort (cython)
31
+ types: [cython]
32
+ - id: isort
33
+ name: isort (pyi)
34
+ types: [pyi]
35
+ - repo: https://github.com/psf/black
36
+ rev: 22.3.0
37
+ hooks:
38
+ - id: black
39
+ args: [--config=pyproject.toml]
40
+ - repo: https://gitlab.com/pycqa/flake8
41
+ rev: 4.0.1
42
+ hooks:
43
+ - id: flake8
44
+ args: [--ignore, "E203", --max-line-length, "79"]
45
+ - repo: https://github.com/kynan/nbstripout
46
+ rev: 0.5.0
47
+ hooks:
48
+ - id: nbstripout
49
+ - repo: https://github.com/asottile/pyupgrade
50
+ rev: v2.32.0
51
+ hooks:
52
+ - id: pyupgrade
53
+ args: [--py36-plus]
.pre-commit-setting.toml ADDED
@@ -0,0 +1,22 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # See https://pre-commit.com for more information
2
+ # See https://pre-commit.com/hooks.html for more hooks
3
+ [tool.black]
4
+ line-length = 79
5
+ include = '\.pyi?$'
6
+ exclude = '''
7
+ /(
8
+ \.git
9
+ | \.idea
10
+ | \.pytest_cache
11
+ | \.tox
12
+ | \.venv
13
+ | _build
14
+ | buck-out
15
+ | build
16
+ | dist
17
+ )/
18
+ '''
19
+
20
+ [flake8]
21
+ ignore = E203
22
+ max-line-length = 79
_config.yml ADDED
@@ -0,0 +1 @@
 
 
1
+ theme: jekyll-theme-modernist
app.py ADDED
@@ -0,0 +1,68 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import gradio as gr
2
+ import openvino as ov
3
+ from pathlib import Path
4
+ import numpy as np
5
+ from PIL import Image
6
+ import cv2
7
+
8
+ from src.config import (
9
+ DICT_DIR,
10
+ IMAGE_TYPES,
11
+ IMAGE_EXAMPLE,
12
+ MODEL_DIR,
13
+ DEVICE,
14
+ )
15
+
16
+ from src.image_processing import recognize
17
+
18
+ # Load models
19
+ core = ov.Core()
20
+ model = core.read_model(model=Path(MODEL_DIR))
21
+ print("[INFO] Loaded recognition model")
22
+
23
+ # Select device (CPU or GPU)
24
+ compiled_model = core.compile_model(model=model, device_name=DEVICE)
25
+
26
+ # Fetch Information About Input and Output Layers
27
+ recognition_input_layer = compiled_model.input(0)
28
+ recognition_output_layer = compiled_model.output(0)
29
+
30
+ print("[INFO] Fetched recognition model")
31
+
32
+ # In JA model, there should be blank symbol added at index 0 of each charlist.
33
+ blank_char = "~"
34
+
35
+ with Path(DICT_DIR).open(mode="r", encoding="utf-8") as charlist:
36
+ letters = blank_char + "".join(line.strip() for line in charlist)
37
+ print("[INFO] Loaded dictionary")
38
+
39
+
40
+
41
+ def do_ocr(inp):
42
+ #img = Image.open(inp).convert('L')
43
+ #img = np.array(img)
44
+ print(f"input: {inp}")
45
+ print(type(inp))
46
+ #img = cv2.imread(inp, cv2.IMREAD_GRAYSCALE)
47
+ img = cv2.cvtColor(inp, cv2.COLOR_BGR2GRAY)
48
+ recognized_text = recognize(img,
49
+ compiled_model,
50
+ recognition_input_layer,
51
+ recognition_output_layer,
52
+ letters,
53
+ )
54
+ return "".join(recognized_text)
55
+
56
+ input = gr.Image()
57
+ output = gr.Textbox()
58
+
59
+ title = "日本語手書き認識"
60
+ description = "DEMO by Tokyo Teachies (注意：画像には1行のテキストしか含まれていません。)"
61
+ examples=[['data/in_1.png'],['data/sample_1_1.png']]
62
+
63
+ gr.Interface(fn=do_ocr,
64
+ inputs=input,
65
+ outputs=output,
66
+ title=title,
67
+ description=description,
68
+ examples=examples).launch()
app_streamlit.py ADDED
@@ -0,0 +1,104 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import streamlit as st
2
+ import openvino as ov
3
+ from pathlib import Path
4
+ import numpy as np
5
+ from PIL import Image
6
+
7
+ from src.config import (
8
+ DICT_DIR,
9
+ IMAGE_TYPES,
10
+ IMAGE_EXAMPLE,
11
+ MODEL_DIR,
12
+ DEVICE,
13
+ )
14
+
15
+ from src.image_processing import recognize
16
+
17
+ # hide deprication warnings which directly don't affect the working of the application
18
+ import warnings
19
+ warnings.filterwarnings("ignore")
20
+
21
+ # Setting custom Page Title and Icon with changed layout and sidebar state
22
+ st.set_page_config(
23
+ page_title="日本語手書き認識",
24
+ page_icon="",
25
+ layout="centered",
26
+ initial_sidebar_state="auto",
27
+ )
28
+
29
+ # hide the part of the code, as this is just for adding some custom CSS styling but not a part of the main idea
30
+ hide_streamlit_style = """
31
+ <style>
32
+ #MainMenu {visibility: hidden;}
33
+ footer {visibility: hidden;}
34
+ </style>
35
+ """
36
+ st.markdown(hide_streamlit_style, unsafe_allow_html=True) # hide the CSS code from the screen as they are embedded in markdown text. Also, allow streamlit to unsafely process as HTML
37
+
38
+
39
+ @st.cache_resource
40
+ def init():
41
+ # Load models
42
+ core = ov.Core()
43
+ model = core.read_model(model=Path(MODEL_DIR))
44
+ print("[INFO] Loaded recognition model")
45
+
46
+ # Select device (CPU or GPU)
47
+ compiled_model = core.compile_model(model=model, device_name=DEVICE)
48
+
49
+ # Fetch Information About Input and Output Layers
50
+ recognition_input_layer = compiled_model.input(0)
51
+ recognition_output_layer = compiled_model.output(0)
52
+
53
+ print("[INFO] Fetched recognition model")
54
+
55
+ # In JA model, there should be blank symbol added at index 0 of each charlist.
56
+ blank_char = "~"
57
+
58
+ with Path(DICT_DIR).open(mode="r", encoding="utf-8") as charlist:
59
+ letters = blank_char + "".join(line.strip() for line in charlist)
60
+ print("[INFO] Loaded dictionary")
61
+
62
+
63
+ return [compiled_model, recognition_input_layer, recognition_output_layer, letters]
64
+
65
+ def display_text(bounds):
66
+ text = []
67
+ for x in bounds:
68
+ t = x[1]
69
+ text.append(t)
70
+ text = ' '.join(text)
71
+ return text
72
+
73
+ HWRmodel = init()
74
+ st.set_option('deprecation.showfileUploaderEncoding',False)
75
+ st.title('日本語手書き認識')
76
+ st.subheader('Tokyo Teachies (DEMO)')
77
+ st.subheader('注意：画像には1行のテキストしか含まれていません。')
78
+ #st.text('Select source Language from the Sidebar.')
79
+
80
+ image_file = st.file_uploader("画像をアップロード…",type=IMAGE_TYPES)
81
+ if image_file is not None:
82
+ st.subheader('アップロードした画像')
83
+ st.image(image_file,width=450)
84
+ else:
85
+ st.subheader('例：')
86
+ image_file = IMAGE_EXAMPLE
87
+ st.image(image_file, use_column_width=450)
88
+
89
+
90
+ if st.button("Recognize"):
91
+ if image_file is not None:
92
+ img = Image.open(image_file).convert('L')
93
+ img = np.array(img)
94
+
95
+ with st.spinner('テキストの抽出...'):
96
+ recognized_text = recognize(img, HWRmodel[0], HWRmodel[1], HWRmodel[2], HWRmodel[3])
97
+ #st.subheader('Extracted text is ...')
98
+ #text = display_text(recognized_text)
99
+ st.write("".join(recognized_text))
100
+
101
+
102
+
103
+
104
+
data/in_1.png ADDED
data/sample_1_1.png ADDED
dict/japanese_charlist.txt ADDED
@@ -0,0 +1,4441 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ 耗
2
+ 雛
3
+ 」
4
+ 輝
5
+ 郎
6
+ 棗
7
+ 寺
8
+ 行
9
+ 〕
10
+ 康
11
+ 藩
12
+ 陀
13
+ 記
14
+ 忙
15
+ 鱈
16
+ 芽
17
+ 詈
18
+ 従
19
+ 攻
20
+ 債
21
+ 戯
22
+ 欅
23
+ 俗
24
+ 郡
25
+ 垣
26
+ 畑
27
+ 次
28
+ 潮
29
+ 駐
30
+ ｇ
31
+ 晟
32
+ 寝
33
+ 砥
34
+ 熹
35
+ 壁
36
+ 齋
37
+ ビ
38
+ 溜
39
+ 蘊
40
+ 舵
41
+ 薊
42
+ 睦
43
+ 登
44
+ 屁
45
+ 堤
46
+ 咎
47
+ 潤
48
+ 栽
49
+ 草
50
+ 粘
51
+ 曰
52
+ 淌
53
+ 佳
54
+ 冤
55
+ 筌
56
+ 梱
57
+ 旺
58
+ 馴
59
+ 蓼
60
+ 鑽
61
+ 涌
62
+ 叉
63
+ 類
64
+ 挑
65
+ 羲
66
+ 設
67
+ 株
68
+ 摩
69
+ 跳
70
+ 科
71
+ 連
72
+ 覚
73
+ 伝
74
+ 逓
75
+ 枳
76
+ 桂
77
+ 迭
78
+ 鯏
79
+ 講
80
+ 繍
81
+ 蕃
82
+ 摯
83
+ 疇
84
+ 覽
85
+ 陽
86
+ 黙
87
+ 鞆
88
+ 革
89
+ 氷
90
+ Ｘ
91
+ 邊
92
+ 逸
93
+ 煎
94
+ 棒
95
+ 甑
96
+ 励
97
+ 鉋
98
+ 罷
99
+ 騫
100
+ 乢
101
+ 院
102
+ 抄
103
+ 漬
104
+ 服
105
+ 蔽
106
+ 惰
107
+ 琺
108
+ 麟
109
+ ｐ
110
+ 思
111
+ 赴
112
+ 邦
113
+ 耿
114
+ 虫
115
+ 鍵
116
+ 資
117
+ 瘁
118
+ 鵺
119
+ 岾
120
+ 煙
121
+ 筒
122
+ 紆
123
+ 褐
124
+ モ
125
+ 、
126
+ ｝
127
+ 更
128
+ 凍
129
+ 宕
130
+ 苅
131
+ 壬
132
+ 痒
133
+ 錚
134
+ 躁
135
+ 外
136
+ 舶
137
+ 羞
138
+ 諮
139
+ 薙
140
+ 橫
141
+ ∨
142
+ 秦
143
+ 倹
144
+ 篠
145
+ ο
146
+ 粒
147
+ た
148
+ 亙
149
+ 非
150
+ 腸
151
+ 樗
152
+ 鐡
153
+ 経
154
+ ユ
155
+ 諺
156
+ 杢
157
+ 歪
158
+ 讚
159
+ 叡
160
+ 倚
161
+ 漱
162
+ 曳
163
+ 偽
164
+ 猯
165
+ 臥
166
+ 珀
167
+ 捐
168
+ 鏡
169
+ 皐
170
+ 既
171
+ 勇
172
+ 寥
173
+ 癈
174
+ 黎
175
+ 栗
176
+ 北
177
+ 吟
178
+ 震
179
+ 腰
180
+ 肆
181
+ 敗
182
+ 泳
183
+ 窪
184
+ 史
185
+ 攫
186
+ 妙
187
+ 叢
188
+ レ
189
+ 諜
190
+ 搜
191
+ 檮
192
+ 秬
193
+ 箙
194
+ 各
195
+ 親
196
+ 四
197
+ 督
198
+ 響
199
+ 硲
200
+ 楊
201
+ 起
202
+ 器
203
+ 承
204
+ 乎
205
+ 寛
206
+ 蘆
207
+ 鉾
208
+ 敢
209
+ 纏
210
+ 擱
211
+ 砕
212
+ 臘
213
+ な
214
+ 淋
215
+ わ
216
+ 憶
217
+ 間
218
+ が
219
+ 洗
220
+ 落
221
+ 蟷
222
+ 収
223
+ ｚ
224
+ 魔
225
+ 輳
226
+ 彩
227
+ 采
228
+ 磑
229
+ 礪
230
+ 受
231
+ 聚
232
+ ⑥
233
+ 秀
234
+ 確
235
+ 険
236
+ 略
237
+ 葛
238
+ 粂
239
+ 啀
240
+ 拮
241
+ 玄
242
+ 幟
243
+ き
244
+ 垂
245
+ ≡
246
+ 瀝
247
+ 榧
248
+ 「
249
+ 絞
250
+ 鞅
251
+ 萌
252
+ 暖
253
+ 悶
254
+ 症
255
+ 几
256
+ 烙
257
+ 埜
258
+ 時
259
+ 働
260
+ 洋
261
+ 賭
262
+ 甸
263
+ 寵
264
+ 葡
265
+ 稀
266
+ 恩
267
+ 銀
268
+ 世
269
+ 剖
270
+ 簗
271
+ 蜑
272
+ 鵯
273
+ 彼
274
+ 禿
275
+ 還
276
+ Ｌ
277
+ 茎
278
+ 焦
279
+ 伽
280
+ 旋
281
+ 熏
282
+ 縊
283
+ 茱
284
+ 濕
285
+ 泰
286
+ 無
287
+ 帆
288
+ 樫
289
+ ェ
290
+ 掬
291
+ 蓬
292
+ 睫
293
+ 駝
294
+ 境
295
+ 竄
296
+ 桁
297
+ 熨
298
+ 客
299
+ ぶ
300
+ 。
301
+ 饋
302
+ へ
303
+ 廳
304
+ 競
305
+ 奬
306
+ ￥
307
+ 梶
308
+ 窕
309
+ 走
310
+ 核
311
+ 付
312
+ 総
313
+ 肺
314
+ 壜
315
+ 將
316
+ 汝
317
+ 顎
318
+ 々
319
+ 〈
320
+ 泉
321
+ 藏
322
+ 篆
323
+ 囲
324
+ 鮖
325
+ 贔
326
+ 筥
327
+ 序
328
+ 亮
329
+ 耶
330
+ 苞
331
+ 怙
332
+ 萢
333
+ 叩
334
+ 噂
335
+ 錨
336
+ 俯
337
+ 問
338
+ 臼
339
+ 瞥
340
+ 輿
341
+ 硬
342
+ 難
343
+ 縄
344
+ ⑨
345
+ 坦
346
+ 帝
347
+ 圓
348
+ 雑
349
+ 得
350
+ 町
351
+ ＝
352
+ 竿
353
+ 潭
354
+ 釡
355
+ 鋪
356
+ Ｇ
357
+ 杵
358
+ 轉
359
+ 杠
360
+ 憮
361
+ 緑
362
+ 4
363
+ 膿
364
+ 慕
365
+ 姫
366
+ 俚
367
+ 便
368
+ 誑
369
+ 看
370
+ 俄
371
+ 炸
372
+ 辜
373
+ 石
374
+ 溯
375
+ 伺
376
+ 覇
377
+ 云
378
+ 陥
379
+ 篇
380
+ 窯
381
+ 效
382
+ 棧
383
+ 蔡
384
+ 銹
385
+ 政
386
+ 館
387
+ ピ
388
+ 味
389
+ 粍
390
+ 旭
391
+ 冒
392
+ 鯨
393
+ 埖
394
+ 卯
395
+ 壱
396
+ 氓
397
+ カ
398
+ 嬉
399
+ Ｂ
400
+ ＿
401
+ 鷆
402
+ 踏
403
+ 懲
404
+ 紳
405
+ 盟
406
+ 蜂
407
+ 堵
408
+ ④
409
+ 鷄
410
+ ヰ
411
+ 胆
412
+ 抱
413
+ 抓
414
+ ＜
415
+ 痺
416
+ 退
417
+ 隅
418
+ 充
419
+ 湫
420
+ 謬
421
+ 電
422
+ 篤
423
+ 卸
424
+ 妖
425
+ 恣
426
+ 裟
427
+ 罫
428
+ 門
429
+ 喘
430
+ ψ
431
+ 聘
432
+ は
433
+ 浅
434
+ 舟
435
+ 礎
436
+ 衷
437
+ 甘
438
+ 倦
439
+ 椒
440
+ 比
441
+ 菜
442
+ 松
443
+ 獣
444
+ 溢
445
+ 鰻
446
+ 肋
447
+ チ
448
+ 避
449
+ 雉
450
+ 饉
451
+ 袱
452
+ 紡
453
+ 笙
454
+ 安
455
+ 鯵
456
+ 参
457
+ 猥
458
+ 目
459
+ 咳
460
+ 垢
461
+ 御
462
+ 眩
463
+ 拾
464
+ 化
465
+ 指
466
+ 淆
467
+ 伎
468
+ 湿
469
+ ｛
470
+ 机
471
+ 浜
472
+ 陵
473
+ 崔
474
+ 噴
475
+ ）
476
+ 殊
477
+ 聊
478
+ 趾
479
+ 礁
480
+ ±
481
+ 泄
482
+ 趙
483
+ 辟
484
+ 刻
485
+ 湾
486
+ ぬ
487
+ 佃
488
+ 緬
489
+ 祭
490
+ 異
491
+ 龍
492
+ 插
493
+ 殲
494
+ 倶
495
+ 似
496
+ 簿
497
+ 桝
498
+ ρ
499
+ 悸
500
+ 殻
501
+ 停
502
+ 蔑
503
+ 洒
504
+ 飯
505
+ 蛯
506
+ 椹
507
+ 争
508
+ 綜
509
+ 螢
510
+ 欧
511
+ ｋ
512
+ あ
513
+ 犠
514
+ 鹸
515
+ 融
516
+ 礫
517
+ ③
518
+ 慌
519
+ 故
520
+ 屑
521
+ 溝
522
+ 岱
523
+ 哩
524
+ 廉
525
+ 鵜
526
+ 蜘
527
+ 羨
528
+ 狡
529
+ 汰
530
+ 瀬
531
+ 1
532
+ 暑
533
+ 酬
534
+ 続
535
+ 禅
536
+ 釉
537
+ 罐
538
+ 騒
539
+ 毀
540
+ 敵
541
+ 騨
542
+ 正
543
+ 棲
544
+ 訝
545
+ 鶇
546
+ 軒
547
+ 碑
548
+ 諭
549
+ ゴ
550
+ 崚
551
+ 惧
552
+ 猴
553
+ 濤
554
+ 鯛
555
+ 瞞
556
+ 継
557
+ 拐
558
+ 痔
559
+ 作
560
+ 交
561
+ 荼
562
+ 剛
563
+ 怖
564
+ 轤
565
+ 牛
566
+ 丑
567
+ 忘
568
+ 斧
569
+ 紛
570
+ 併
571
+ ｄ
572
+ ｈ
573
+ －
574
+ 篁
575
+ 鎭
576
+ 匂
577
+ ネ
578
+ 加
579
+ 贈
580
+ 潴
581
+ 皓
582
+ 超
583
+ 桴
584
+ 爆
585
+ 嚴
586
+ 菫
587
+ 竜
588
+ 仮
589
+ 徴
590
+ ｖ
591
+ 塙
592
+ 席
593
+ 背
594
+ 伸
595
+ 醇
596
+ 繩
597
+ 尉
598
+ 倣
599
+ 五
600
+ 壮
601
+ 緩
602
+ 槻
603
+ 邸
604
+ 樒
605
+ 鵲
606
+ 釛
607
+ 辛
608
+ 徊
609
+ 跿
610
+ 験
611
+ 直
612
+ メ
613
+ 平
614
+ 雪
615
+ 拉
616
+ 辷
617
+ ［
618
+ 冶
619
+ 風
620
+ 笥
621
+ 打
622
+ 虻
623
+ 朽
624
+ 珈
625
+ 爭
626
+ 臓
627
+ 綠
628
+ 冊
629
+ 柘
630
+ 靄
631
+ 居
632
+ 演
633
+ λ
634
+ 刺
635
+ 懇
636
+ 恒
637
+ 澹
638
+ 寳
639
+ 窮
640
+ 智
641
+ 椦
642
+ 負
643
+ 園
644
+ 薑
645
+ 漢
646
+ 拍
647
+ 樋
648
+ 夢
649
+ 雨
650
+ 莊
651
+ 痰
652
+ π
653
+ 椣
654
+ 漠
655
+ 莇
656
+ 種
657
+ 詰
658
+ 巧
659
+ 替
660
+ 悴
661
+ ー
662
+ 戔
663
+ 揖
664
+ 柧
665
+ 賂
666
+ 恤
667
+ 復
668
+ 宍
669
+ 愁
670
+ 謹
671
+ 色
672
+ 契
673
+ ×
674
+ 眼
675
+ 芋
676
+ 柱
677
+ 構
678
+ 塑
679
+ 寂
680
+ 宜
681
+ 胎
682
+ 謀
683
+ 齊
684
+ 郞
685
+ 賊
686
+ 櫨
687
+ 阿
688
+ 斃
689
+ 臍
690
+ 匙
691
+ 吸
692
+ 肖
693
+ 結
694
+ 誕
695
+ 礒
696
+ 顰
697
+ 鳰
698
+ 崇
699
+ 決
700
+ 僅
701
+ 蝶
702
+ 典
703
+ 心
704
+ 節
705
+ 籤
706
+ 腑
707
+ ∴
708
+ ス
709
+ 壯
710
+ 悍
711
+ 沮
712
+ 矛
713
+ ‥
714
+ 据
715
+ 氣
716
+ 租
717
+ 鄭
718
+ 熔
719
+ 臾
720
+ 巳
721
+ 詩
722
+ 蓚
723
+ 役
724
+ 諧
725
+ 援
726
+ 縛
727
+ 揚
728
+ 犯
729
+ 夕
730
+ 貰
731
+ 苑
732
+ 呈
733
+ 炒
734
+ κ
735
+ 盆
736
+ 或
737
+ ぷ
738
+ 岬
739
+ 壕
740
+ 弐
741
+ 已
742
+ 睾
743
+ 糧
744
+ 愾
745
+ 胴
746
+ 琵
747
+ 候
748
+ 撤
749
+ 桜
750
+ 鼻
751
+ 涼
752
+ 諷
753
+ 叔
754
+ 央
755
+ 概
756
+ 歿
757
+ 皷
758
+ 戰
759
+ 硝
760
+ 蛍
761
+ 姥
762
+ 專
763
+ 内
764
+ 莨
765
+ 侠
766
+ 杭
767
+ 芹
768
+ 麓
769
+ 杙
770
+ 謂
771
+ 製
772
+ 嘘
773
+ 禪
774
+ 休
775
+ 涜
776
+ ぱ
777
+ 泪
778
+ 升
779
+ 襲
780
+ 二
781
+ 釦
782
+ 卷
783
+ 鮹
784
+ 烈
785
+ ↓
786
+ 僕
787
+ 額
788
+ ⑩
789
+ 盗
790
+ 彷
791
+ 玖
792
+ 汽
793
+ 先
794
+ 描
795
+ 耐
796
+ 柁
797
+ 》
798
+ 錯
799
+ 隠
800
+ 授
801
+ Ａ
802
+ 斡
803
+ 噺
804
+ 樅
805
+ 妬
806
+ 鮠
807
+ 池
808
+ Ｅ
809
+ 峡
810
+ 沢
811
+ 拶
812
+ 蘇
813
+ 暘
814
+ 派
815
+ 荊
816
+ ザ
817
+ 什
818
+ 渚
819
+ 紀
820
+ 稚
821
+ 傭
822
+ ヂ
823
+ 輯
824
+ 疵
825
+ 浣
826
+ 怨
827
+ 跡
828
+ 被
829
+ 球
830
+ 槁
831
+ 躰
832
+ 盲
833
+ 束
834
+ 閘
835
+ 卿
836
+ 隈
837
+ 来
838
+ 其
839
+ 簡
840
+ 梯
841
+ 薮
842
+ 折
843
+ 苺
844
+ 栞
845
+ 貶
846
+ 董
847
+ 虐
848
+ 警
849
+ 準
850
+ 億
851
+ 沐
852
+ ブ
853
+ 染
854
+ 牒
855
+ 題
856
+ 鯉
857
+ 舗
858
+ 狂
859
+ 守
860
+ 袖
861
+ 格
862
+ Ⅴ
863
+ 蓋
864
+ 宋
865
+ 涛
866
+ 睛
867
+ 蹲
868
+ 崩
869
+ 勤
870
+ 了
871
+ ヮ
872
+ 焙
873
+ 介
874
+ ゐ
875
+ 惟
876
+ 弦
877
+ 稲
878
+ 桧
879
+ 鵡
880
+ 0
881
+ 臭
882
+ 弼
883
+ 瞠
884
+ 赦
885
+ 戊
886
+ 朱
887
+ 啜
888
+ 剃
889
+ 瑚
890
+ 圀
891
+ 示
892
+ 常
893
+ 穆
894
+ 訃
895
+ 筋
896
+ 頬
897
+ 鮴
898
+ 誓
899
+ 樣
900
+ 劍
901
+ 騰
902
+ 梨
903
+ 蠅
904
+ 廣
905
+ 逢
906
+ 橘
907
+ 月
908
+ 庁
909
+ 乍
910
+ 匡
911
+ と
912
+ 晋
913
+ 喉
914
+ 麺
915
+ リ
916
+ 存
917
+ 婪
918
+ 銚
919
+ 伊
920
+ 豊
921
+ 孑
922
+ 與
923
+ 衾
924
+ 穫
925
+ 捜
926
+ 定
927
+ 編
928
+ 鵠
929
+ 痘
930
+ 崗
931
+ 網
932
+ 声
933
+ 靴
934
+ ヱ
935
+ 仄
936
+ 醐
937
+ 催
938
+ 泓
939
+ 衙
940
+ 蜷
941
+ 誂
942
+ 憐
943
+ 追
944
+ 織
945
+ 塔
946
+ 蚤
947
+ 盾
948
+ 移
949
+ 恢
950
+ 瑣
951
+ 逃
952
+ 引
953
+ 仁
954
+ 搬
955
+ 硼
956
+ 掠
957
+ 術
958
+ 毎
959
+ 燎
960
+ Φ
961
+ 蟇
962
+ 嗜
963
+ 錬
964
+ 綴
965
+ 方
966
+ 預
967
+ 梹
968
+ 原
969
+ 柳
970
+ 虹
971
+ 鯰
972
+ 分
973
+ 踞
974
+ 週
975
+ び
976
+ 杓
977
+ 註
978
+ 襦
979
+ 栄
980
+ 喪
981
+ 慈
982
+ 團
983
+ 站
984
+ 簸
985
+ 聟
986
+ ％
987
+ 多
988
+ 伏
989
+ 鴎
990
+ 遁
991
+ 糞
992
+ 躬
993
+ 仇
994
+ 鉤
995
+ 過
996
+ 惡
997
+ ほ
998
+ 毛
999
+ 凄
1000
+ 貼
1001
+ 爲
1002
+ 霊
1003
+ ｔ
1004
+ Ｈ
1005
+ 六
1006
+ 官
1007
+ 咲
1008
+ 刃
1009
+ 煤
1010
+ 眈
1011
+ 徐
1012
+ 澳
1013
+ ジ
1014
+ 啼
1015
+ 潔
1016
+ 膜
1017
+ 小
1018
+ 珍
1019
+ 習
1020
+ ヨ
1021
+ 怯
1022
+ ァ
1023
+ 稟
1024
+ 諦
1025
+ 睿
1026
+ 貞
1027
+ 締
1028
+ 妨
1029
+ 寸
1030
+ ツ
1031
+ 押
1032
+ ｗ
1033
+ ⑰
1034
+ 彰
1035
+ 洌
1036
+ 誦
1037
+ 暇
1038
+ 鰄
1039
+ 傍
1040
+ 偉
1041
+ 穐
1042
+ 集
1043
+ ゼ
1044
+ 3
1045
+ 鰕
1046
+ ィ
1047
+ 蓮
1048
+ 銅
1049
+ 恬
1050
+ 般
1051
+ ま
1052
+ 祈
1053
+ 炭
1054
+ 都
1055
+ 騎
1056
+ 樂
1057
+ 蒙
1058
+ 選
1059
+ 勉
1060
+ 動
1061
+ 勃
1062
+ 圭
1063
+ 髙
1064
+ 肥
1065
+ 栫
1066
+ 獰
1067
+ 図
1068
+ 韻
1069
+ ゎ
1070
+ ソ
1071
+ 祚
1072
+ 浪
1073
+ 汚
1074
+ ＠
1075
+ 箸
1076
+ 閲
1077
+ 春
1078
+ 巨
1079
+ 紬
1080
+ 繋
1081
+ 牌
1082
+ 芳
1083
+ 蛎
1084
+ 膝
1085
+ 皀
1086
+ 拱
1087
+ 疾
1088
+ 伐
1089
+ 転
1090
+ 陣
1091
+ 罹
1092
+ 広
1093
+ 穂
1094
+ 狭
1095
+ 飢
1096
+ 蹴
1097
+ 粗
1098
+ 凸
1099
+ 遼
1100
+ 幹
1101
+ 忍
1102
+ 桐
1103
+ 律
1104
+ 衛
1105
+ （
1106
+ 葬
1107
+ 隼
1108
+ 玻
1109
+ 鍜
1110
+ 楼
1111
+ 滞
1112
+ 花
1113
+ 冨
1114
+ 業
1115
+ 梃
1116
+ 錆
1117
+ 様
1118
+ 楪
1119
+ 鸚
1120
+ 藤
1121
+ 三
1122
+ 廩
1123
+ 糺
1124
+ 違
1125
+ 初
1126
+ 渭
1127
+ 籍
1128
+ 桑
1129
+ 苓
1130
+ 吹
1131
+ 県
1132
+ 錮
1133
+ 鞠
1134
+ 訥
1135
+ 波
1136
+ 流
1137
+ 仗
1138
+ 奇
1139
+ 俺
1140
+ ヤ
1141
+ 汗
1142
+ 糜
1143
+ 独
1144
+ ョ
1145
+ 庭
1146
+ 陰
1147
+ 蕪
1148
+ 楽
1149
+ ｌ
1150
+ 掛
1151
+ 子
1152
+ 祗
1153
+ い
1154
+ 錫
1155
+ 群
1156
+ 俎
1157
+ 俤
1158
+ 呑
1159
+ 使
1160
+ 糖
1161
+ 籾
1162
+ 凱
1163
+ 鍛
1164
+ 脹
1165
+ 淡
1166
+ 耽
1167
+ 傘
1168
+ 鰍
1169
+ 唄
1170
+ 協
1171
+ 扣
1172
+ 募
1173
+ 下
1174
+ 痴
1175
+ ふ
1176
+ 瑕
1177
+ 翫
1178
+ 卵
1179
+ 槌
1180
+ 繊
1181
+ ぎ
1182
+ 腱
1183
+ ε
1184
+ 簑
1185
+ 枚
1186
+ 寮
1187
+ ろ
1188
+ 依
1189
+ 抹
1190
+ 浩
1191
+ 梓
1192
+ ポ
1193
+ 琉
1194
+ 梧
1195
+ 抑
1196
+ 土
1197
+ 伯
1198
+ 枯
1199
+ 墾
1200
+ 蝕
1201
+ 推
1202
+ 雄
1203
+ 蹠
1204
+ 息
1205
+ θ
1206
+ 蛇
1207
+ 玲
1208
+ 籔
1209
+ 駒
1210
+ 莫
1211
+ 羊
1212
+ 仲
1213
+ 渡
1214
+ ＃
1215
+ 容
1216
+ 歯
1217
+ 滅
1218
+ 鰌
1219
+ 載
1220
+ 蛹
1221
+ 卒
1222
+ フ
1223
+ 鳶
1224
+ 嫉
1225
+ 楢
1226
+ 畷
1227
+ 斜
1228
+ 敍
1229
+ 宝
1230
+ 燦
1231
+ 戮
1232
+ ｘ
1233
+ 終
1234
+ 双
1235
+ 儉
1236
+ 粭
1237
+ 軋
1238
+ 撹
1239
+ 留
1240
+ 裳
1241
+ 蒸
1242
+ 凹
1243
+ 彌
1244
+ 改
1245
+ 並
1246
+ 森
1247
+ 輸
1248
+ 釜
1249
+ 肛
1250
+ 沽
1251
+ 灘
1252
+ 借
1253
+ 前
1254
+ 搾
1255
+ 生
1256
+ 奏
1257
+ 備
1258
+ 憩
1259
+ 喚
1260
+ 眺
1261
+ 吝
1262
+ 緘
1263
+ 2
1264
+ 妊
1265
+ 胞
1266
+ 己
1267
+ 学
1268
+ 越
1269
+ 柿
1270
+ 華
1271
+ 莞
1272
+ 顕
1273
+ 拷
1274
+ 捉
1275
+ 至
1276
+ 妓
1277
+ 篦
1278
+ 陷
1279
+ 凡
1280
+ 粁
1281
+ 昔
1282
+ お
1283
+ Ｑ
1284
+ 馨
1285
+ 躑
1286
+ 膀
1287
+ ∥
1288
+ 能
1289
+ 涙
1290
+ 狼
1291
+ 蕁
1292
+ 犀
1293
+ 通
1294
+ 田
1295
+ ⑤
1296
+ 濁
1297
+ 昭
1298
+ 郵
1299
+ 架
1300
+ 淨
1301
+ 淫
1302
+ 盜
1303
+ 收
1304
+ 産
1305
+ 7
1306
+ 循
1307
+ 痩
1308
+ 橇
1309
+ 莚
1310
+ 猶
1311
+ 蔗
1312
+ 級
1313
+ 郛
1314
+ 挨
1315
+ 勅
1316
+ 咽
1317
+ 羅
1318
+ 重
1319
+ 占
1320
+ 隘
1321
+ 失
1322
+ 粋
1323
+ 趣
1324
+ 鰯
1325
+ 嵩
1326
+ 吻
1327
+ れ
1328
+ 佇
1329
+ 悽
1330
+ 麾
1331
+ 咄
1332
+ 聞
1333
+ 緤
1334
+ 棟
1335
+ 奘
1336
+ 苧
1337
+ 駿
1338
+ 抬
1339
+ 艮
1340
+ 瞑
1341
+ 徽
1342
+ Ｃ
1343
+ 沌
1344
+ 慣
1345
+ 頒
1346
+ 錺
1347
+ 績
1348
+ 造
1349
+ 鷺
1350
+ 彙
1351
+ 檢
1352
+ 餅
1353
+ せ
1354
+ 談
1355
+ 刷
1356
+ 薗
1357
+ 丘
1358
+ 倖
1359
+ 操
1360
+ 納
1361
+ 墹
1362
+ 剪
1363
+ 犬
1364
+ 嶬
1365
+ 瓏
1366
+ 脱
1367
+ 想
1368
+ 轡
1369
+ 歌
1370
+ 秩
1371
+ 虞
1372
+ 忝
1373
+ 證
1374
+ 煥
1375
+ 埴
1376
+ 手
1377
+ 鞭
1378
+ δ
1379
+ 菟
1380
+ 櫻
1381
+ 訊
1382
+ 鏑
1383
+ ※
1384
+ 粕
1385
+ 符
1386
+ 畭
1387
+ 鯆
1388
+ や
1389
+ 萠
1390
+ 剰
1391
+ 殺
1392
+ 汐
1393
+ ｙ
1394
+ 削
1395
+ 末
1396
+ ≒
1397
+ 坪
1398
+ バ
1399
+ →
1400
+ 紅
1401
+ 琶
1402
+ 制
1403
+ 幕
1404
+ ＾
1405
+ 炎
1406
+ 臆
1407
+ 狄
1408
+ 碓
1409
+ 娯
1410
+ 灼
1411
+ 沫
1412
+ 写
1413
+ 寔
1414
+ 肪
1415
+ 堡
1416
+ 祥
1417
+ 責
1418
+ 善
1419
+ 岡
1420
+ 菩
1421
+ 恭
1422
+ 慮
1423
+ 菅
1424
+ 扮
1425
+ 拳
1426
+ 壊
1427
+ 椋
1428
+ 箏
1429
+ 迦
1430
+ 曇
1431
+ 賞
1432
+ 聴
1433
+ 只
1434
+ 榔
1435
+ 欄
1436
+ 料
1437
+ 朝
1438
+ 飜
1439
+ 甦
1440
+ 縢
1441
+ 秋
1442
+ 局
1443
+ 此
1444
+ 綺
1445
+ 茫
1446
+ 脊
1447
+ 弁
1448
+ 穰
1449
+ 湘
1450
+ 腐
1451
+ 瑞
1452
+ 浄
1453
+ 脅
1454
+ 跪
1455
+ 冴
1456
+ 惑
1457
+ 林
1458
+ ω
1459
+ 袁
1460
+ 恆
1461
+ 綱
1462
+ 猷
1463
+ Ⅰ
1464
+ 宅
1465
+ 兆
1466
+ 好
1467
+ 我
1468
+ 暉
1469
+ 扇
1470
+ 搦
1471
+ 慟
1472
+ 蚫
1473
+ 竈
1474
+ 闘
1475
+ 佗
1476
+ 匿
1477
+ 莉
1478
+ 吽
1479
+ 曼
1480
+ 模
1481
+ 遥
1482
+ 莪
1483
+ 抒
1484
+ 劫
1485
+ 階
1486
+ 艀
1487
+ 栂
1488
+ 僧
1489
+ 蝮
1490
+ 章
1491
+ 檻
1492
+ 顛
1493
+ 稽
1494
+ 頸
1495
+ 錘
1496
+ 褪
1497
+ 閖
1498
+ 閉
1499
+ 凶
1500
+ 蝿
1501
+ 葵
1502
+ 斉
1503
+ 批
1504
+ 賦
1505
+ 悼
1506
+ 筆
1507
+ 円
1508
+ 尖
1509
+ 陋
1510
+ 薦
1511
+ 塊
1512
+ 漆
1513
+ 天
1514
+ 握
1515
+ 喧
1516
+ 映
1517
+ 杜
1518
+ 捲
1519
+ 米
1520
+ 乘
1521
+ 報
1522
+ 丁
1523
+ 李
1524
+ 嶮
1525
+ 悉
1526
+ 墜
1527
+ 塀
1528
+ 嫌
1529
+ 湊
1530
+ 省
1531
+ 野
1532
+ 曽
1533
+ 洪
1534
+ 膨
1535
+ 蛉
1536
+ 九
1537
+ 進
1538
+ 労
1539
+ 統
1540
+ 賠
1541
+ 忖
1542
+ 奔
1543
+ 液
1544
+ Ｚ
1545
+ 縋
1546
+ 罠
1547
+ 陪
1548
+ 諌
1549
+ 條
1550
+ 爽
1551
+ 禰
1552
+ 輻
1553
+ 荒
1554
+ 蓑
1555
+ 摸
1556
+ 茂
1557
+ 栁
1558
+ 頷
1559
+ 坂
1560
+ 際
1561
+ 将
1562
+ 緊
1563
+ 柾
1564
+ 靖
1565
+ 楔
1566
+ 窟
1567
+ 巴
1568
+ 梁
1569
+ 洵
1570
+ 簾
1571
+ 疫
1572
+ 捕
1573
+ 毬
1574
+ 桔
1575
+ 惺
1576
+ 褒
1577
+ 筵
1578
+ 昧
1579
+ 孵
1580
+ 霧
1581
+ 愚
1582
+ 酸
1583
+ 鍍
1584
+ 麹
1585
+ 扉
1586
+ 繭
1587
+ 磋
1588
+ 期
1589
+ 茶
1590
+ 嵐
1591
+ 鋏
1592
+ 自
1593
+ 叱
1594
+ 遺
1595
+ 謁
1596
+ 庇
1597
+ 頃
1598
+ 銑
1599
+ 減
1600
+ 藻
1601
+ 蕾
1602
+ 東
1603
+ 庚
1604
+ 鰐
1605
+ 狹
1606
+ ⑱
1607
+ 尻
1608
+ 饒
1609
+ 誘
1610
+ も
1611
+ 夥
1612
+ 校
1613
+ 曻
1614
+ 列
1615
+ 痕
1616
+ 鴦
1617
+ 斗
1618
+ 共
1619
+ 胃
1620
+ 詑
1621
+ 再
1622
+ 缶
1623
+ 軸
1624
+ 遅
1625
+ 蜻
1626
+ 增
1627
+ 吋
1628
+ 狷
1629
+ 穏
1630
+ 朋
1631
+ 足
1632
+ Ⅳ
1633
+ 蹟
1634
+ 奨
1635
+ 姪
1636
+ 霖
1637
+ 頌
1638
+ 象
1639
+ 眄
1640
+ 貌
1641
+ 噪
1642
+ 瑤
1643
+ 丙
1644
+ 囮
1645
+ 葹
1646
+ 藪
1647
+ 駕
1648
+ 兎
1649
+ 屋
1650
+ 夙
1651
+ 炊
1652
+ 相
1653
+ 婁
1654
+ 姻
1655
+ ざ
1656
+ 藥
1657
+ 及
1658
+ 隷
1659
+ 准
1660
+ 光
1661
+ 蒄
1662
+ 稗
1663
+ 粧
1664
+ 筬
1665
+ Σ
1666
+ 廻
1667
+ 紙
1668
+ 醜
1669
+ 貯
1670
+ 兵
1671
+ 暴
1672
+ 輌
1673
+ 竺
1674
+ 屠
1675
+ 詫
1676
+ 俵
1677
+ 斂
1678
+ 瓦
1679
+ 鑪
1680
+ 規
1681
+ 廼
1682
+ 衆
1683
+ 択
1684
+ 和
1685
+ 年
1686
+ 慾
1687
+ 福
1688
+ 陸
1689
+ 考
1690
+ 満
1691
+ オ
1692
+ 戲
1693
+ ］
1694
+ 欠
1695
+ 堝
1696
+ 舛
1697
+ 鱸
1698
+ 票
1699
+ 樽
1700
+ 任
1701
+ 戎
1702
+ 勲
1703
+ 優
1704
+ 苗
1705
+ 但
1706
+ 怜
1707
+ 賛
1708
+ 嚆
1709
+ 鰹
1710
+ 拘
1711
+ 招
1712
+ 酵
1713
+ 甚
1714
+ 烹
1715
+ 烟
1716
+ ｒ
1717
+ 検
1718
+ 興
1719
+ 榛
1720
+ ÷
1721
+ 乱
1722
+ 牡
1723
+ 鋲
1724
+ 斌
1725
+ 淵
1726
+ 奮
1727
+ 畢
1728
+ 膂
1729
+ 君
1730
+ 巖
1731
+ 鹿
1732
+ 梢
1733
+ 朗
1734
+ 兼
1735
+ 濃
1736
+ 瑶
1737
+ ｏ
1738
+ 垰
1739
+ 則
1740
+ 裝
1741
+ 裔
1742
+ 父
1743
+ 頴
1744
+ 箔
1745
+ 課
1746
+ 程
1747
+ 蒋
1748
+ 閑
1749
+ 篝
1750
+ 深
1751
+ 美
1752
+ 箴
1753
+ 限
1754
+ 聰
1755
+ 以
1756
+ ⑬
1757
+ 萎
1758
+ 盪
1759
+ 閻
1760
+ 憧
1761
+ 橦
1762
+ 堂
1763
+ 瑠
1764
+ 幼
1765
+ 宴
1766
+ 予
1767
+ 汁
1768
+ 燻
1769
+ 颯
1770
+ 劾
1771
+ 絎
1772
+ 眸
1773
+ 佑
1774
+ 挺
1775
+ 形
1776
+ 凧
1777
+ 駄
1778
+ 駮
1779
+ 費
1780
+ ⑧
1781
+ 嬢
1782
+ 蕉
1783
+ 急
1784
+ 岫
1785
+ 皆
1786
+ 離
1787
+ ラ
1788
+ 針
1789
+ 鉱
1790
+ 零
1791
+ 榿
1792
+ 繰
1793
+ 厘
1794
+ 萄
1795
+ 弩
1796
+ 因
1797
+ 澤
1798
+ 汀
1799
+ 訳
1800
+ 倉
1801
+ 衝
1802
+ 姶
1803
+ 絵
1804
+ 拠
1805
+ 罵
1806
+ 鯱
1807
+ 貴
1808
+ 茄
1809
+ 猜
1810
+ 賣
1811
+ 壞
1812
+ 匝
1813
+ 牲
1814
+ Ｔ
1815
+ 紫
1816
+ 耄
1817
+ 洲
1818
+ 拭
1819
+ 餞
1820
+ 悲
1821
+ 沖
1822
+ 麿
1823
+ 導
1824
+ 黒
1825
+ 況
1826
+ 征
1827
+ ヲ
1828
+ 腺
1829
+ 庸
1830
+ 岻
1831
+ 屈
1832
+ 髪
1833
+ 帷
1834
+ 稻
1835
+ ち
1836
+ え
1837
+ 瞭
1838
+ ｑ
1839
+ 髻
1840
+ 王
1841
+ 纈
1842
+ 峠
1843
+ 辣
1844
+ 漂
1845
+ ハ
1846
+ 柵
1847
+ 樵
1848
+ 汞
1849
+ 軌
1850
+ 袋
1851
+ 児
1852
+ 撞
1853
+ 跛
1854
+ 碆
1855
+ 涵
1856
+ 攪
1857
+ 練
1858
+ 寅
1859
+ 蚊
1860
+ 猪
1861
+ 彎
1862
+ 迷
1863
+ 婉
1864
+ 価
1865
+ 揣
1866
+ 累
1867
+ 埃
1868
+ 緻
1869
+ 蛛
1870
+ 開
1871
+ 案
1872
+ 農
1873
+ 媒
1874
+ 娩
1875
+ 呟
1876
+ 塁
1877
+ 絶
1878
+ 凝
1879
+ 鶴
1880
+ 亀
1881
+ 抜
1882
+ 如
1883
+ 股
1884
+ 熾
1885
+ ぺ
1886
+ 窄
1887
+ 防
1888
+ 妹
1889
+ 鎧
1890
+ 没
1891
+ グ
1892
+ 這
1893
+ 煕
1894
+ 執
1895
+ 瑛
1896
+ 蜀
1897
+ 浮
1898
+ 緯
1899
+ 粛
1900
+ 送
1901
+ 桾
1902
+ の
1903
+ 唾
1904
+ 杏
1905
+ 揮
1906
+ 泌
1907
+ 芯
1908
+ 袂
1909
+ 誤
1910
+ ＼
1911
+ ⑯
1912
+ 冠
1913
+ 縺
1914
+ 南
1915
+ 惣
1916
+ 篭
1917
+ 贅
1918
+ 添
1919
+ 鉛
1920
+ 潅
1921
+ 侈
1922
+ 畏
1923
+ 同
1924
+ 劇
1925
+ 箋
1926
+ 杖
1927
+ 歴
1928
+ 培
1929
+ 杁
1930
+ 大
1931
+ 志
1932
+ 祐
1933
+ 熱
1934
+ 筏
1935
+ 芸
1936
+ 坐
1937
+ 飛
1938
+ ケ
1939
+ 済
1940
+ 郊
1941
+ 餠
1942
+ 淑
1943
+ 之
1944
+ 沈
1945
+ 托
1946
+ 逹
1947
+ ｕ
1948
+ 喰
1949
+ 駁
1950
+ 柏
1951
+ 弗
1952
+ 焚
1953
+ 嘩
1954
+ 雁
1955
+ 後
1956
+ ⑭
1957
+ 壽
1958
+ 瞰
1959
+ 鍼
1960
+ 環
1961
+ 止
1962
+ 晦
1963
+ 遡
1964
+ 姑
1965
+ 袴
1966
+ け
1967
+ 敬
1968
+ 拓
1969
+ 署
1970
+ 羽
1971
+ 呰
1972
+ 義
1973
+ 託
1974
+ 逐
1975
+ ズ
1976
+ 俊
1977
+ 碇
1978
+ 泥
1979
+ 唯
1980
+ ロ
1981
+ 衍
1982
+ 侯
1983
+ 利
1984
+ 老
1985
+ 誠
1986
+ 患
1987
+ デ
1988
+ 鮮
1989
+ 坊
1990
+ 随
1991
+ 蓄
1992
+ ≦
1993
+ 彦
1994
+ 呎
1995
+ 慢
1996
+ 蜆
1997
+ 來
1998
+ し
1999
+ 髮
2000
+ 瀧
2001
+ そ
2002
+ 鰈
2003
+ 諒
2004
+ 椰
2005
+ 唐
2006
+ 梍
2007
+ 整
2008
+ 隊
2009
+ 喋
2010
+ 願
2011
+ 淘
2012
+ 菌
2013
+ 巽
2014
+ Ｙ
2015
+ 幌
2016
+ 班
2017
+ 莱
2018
+ 哨
2019
+ 楯
2020
+ 例
2021
+ 脆
2022
+ 渠
2023
+ 遠
2024
+ 鞁
2025
+ 螺
2026
+ 袈
2027
+ 愴
2028
+ 勘
2029
+ 恪
2030
+ 中
2031
+ 室
2032
+ 舞
2033
+ 茜
2034
+ 陦
2035
+ 怠
2036
+ 宣
2037
+ シ
2038
+ 漉
2039
+ 鰔
2040
+ 葱
2041
+ 棯
2042
+ 晴
2043
+ 秒
2044
+ 葎
2045
+ 夜
2046
+ 奄
2047
+ 呪
2048
+ 事
2049
+ 斑
2050
+ ��
2051
+ 眠
2052
+ β
2053
+ 猟
2054
+ 糟
2055
+ 雌
2056
+ 畜
2057
+ 6
2058
+ 庖
2059
+ 獺
2060
+ ゾ
2061
+ 郷
2062
+ 錐
2063
+ 轍
2064
+ 機
2065
+ 晨
2066
+ 碌
2067
+ 孕
2068
+ 畔
2069
+ ど
2070
+ 蝉
2071
+ 塚
2072
+ 召
2073
+ 奧
2074
+ 籃
2075
+ 欽
2076
+ 桒
2077
+ 扁
2078
+ 崖
2079
+ 吐
2080
+ 空
2081
+ 嗚
2082
+ 揺
2083
+ 套
2084
+ 板
2085
+ 鸞
2086
+ 珪
2087
+ 艇
2088
+ 棹
2089
+ 近
2090
+ 媛
2091
+ 嵯
2092
+ 麒
2093
+ 隔
2094
+ 窃
2095
+ Ｋ
2096
+ 影
2097
+ 去
2098
+ 券
2099
+ 穿
2100
+ 麦
2101
+ 蠣
2102
+ τ
2103
+ 言
2104
+ 飴
2105
+ 途
2106
+ 竦
2107
+ ぉ
2108
+ 鮫
2109
+ 曾
2110
+ 瓩
2111
+ 憫
2112
+ 破
2113
+ 捩
2114
+ 精
2115
+ 念
2116
+ 苫
2117
+ 綬
2118
+ 釆
2119
+ 眞
2120
+ 誅
2121
+ 嘲
2122
+ 系
2123
+ 歇
2124
+ 膚
2125
+ 蛋
2126
+ 枡
2127
+ 稼
2128
+ 荘
2129
+ 貢
2130
+ 櫟
2131
+ 鍋
2132
+ 倆
2133
+ 魚
2134
+ 寡
2135
+ 碁
2136
+ 驗
2137
+ 栃
2138
+ 撥
2139
+ 椎
2140
+ 賤
2141
+ 証
2142
+ 翔
2143
+ 童
2144
+ 諾
2145
+ 倪
2146
+ 頁
2147
+ 孤
2148
+ 部
2149
+ 侑
2150
+ 碎
2151
+ 誹
2152
+ 沓
2153
+ 龝
2154
+ 邏
2155
+ 弱
2156
+ 懈
2157
+ 涸
2158
+ ａ
2159
+ 闊
2160
+ 値
2161
+ 恵
2162
+ 路
2163
+ 給
2164
+ 長
2165
+ 鳴
2166
+ 畉
2167
+ 嘗
2168
+ 躍
2169
+ 湛
2170
+ 匪
2171
+ 而
2172
+ 耆
2173
+ 祥
2174
+ 碗
2175
+ 就
2176
+ 謗
2177
+ 孟
2178
+ 瘻
2179
+ 恃
2180
+ 社
2181
+ 尊
2182
+ Ⅲ
2183
+ 捻
2184
+ 狛
2185
+ 蕗
2186
+ 浚
2187
+ 蕩
2188
+ 虜
2189
+ 鰺
2190
+ 溺
2191
+ 頗
2192
+ 挿
2193
+ 貪
2194
+ 膳
2195
+ 瞳
2196
+ 妻
2197
+ ベ
2198
+ 啖
2199
+ 矢
2200
+ ∞
2201
+ 靡
2202
+ 滕
2203
+ ｆ
2204
+ 座
2205
+ 儀
2206
+ 佼
2207
+ 炬
2208
+ 札
2209
+ 蜜
2210
+ 涎
2211
+ 売
2212
+ 悔
2213
+ 裏
2214
+ 党
2215
+ 尋
2216
+ 海
2217
+ 強
2218
+ 委
2219
+ 杞
2220
+ だ
2221
+ 余
2222
+ 傳
2223
+ 粹
2224
+ 〔
2225
+ 藍
2226
+ 昼
2227
+ 幇
2228
+ 黍
2229
+ 疏
2230
+ 躾
2231
+ 應
2232
+ 盛
2233
+ 藷
2234
+ 按
2235
+ 煮
2236
+ 祖
2237
+ 隙
2238
+ 艘
2239
+ 孫
2240
+ 戚
2241
+ 爪
2242
+ じ
2243
+ 策
2244
+ 奸
2245
+ ギ
2246
+ 揉
2247
+ 罅
2248
+ 恙
2249
+ 植
2250
+ 蓉
2251
+ 瑳
2252
+ 壘
2253
+ ｅ
2254
+ 宏
2255
+ 骸
2256
+ 商
2257
+ 疹
2258
+ 磐
2259
+ 何
2260
+ 殖
2261
+ 逝
2262
+ 悛
2263
+ 詢
2264
+ 釧
2265
+ 攬
2266
+ 5
2267
+ 祉
2268
+ 檄
2269
+ に
2270
+ 迎
2271
+ ヵ
2272
+ 琥
2273
+ 朔
2274
+ 江
2275
+ 果
2276
+ 張
2277
+ 謫
2278
+ 永
2279
+ 譲
2280
+ 寢
2281
+ 鑄
2282
+ 論
2283
+ 敏
2284
+ 区
2285
+ ヅ
2286
+ 蟹
2287
+ 喫
2288
+ 聾
2289
+ ガ
2290
+ 癇
2291
+ 衰
2292
+ 劔
2293
+ 認
2294
+ 糎
2295
+ 庫
2296
+ 個
2297
+ 装
2298
+ 饑
2299
+ 蒔
2300
+ ’
2301
+ 惚
2302
+ 干
2303
+ 水
2304
+ 燥
2305
+ 贖
2306
+ 懷
2307
+ 刈
2308
+ 姿
2309
+ 肉
2310
+ 鉄
2311
+ 亭
2312
+ 主
2313
+ 邉
2314
+ 縁
2315
+ 栩
2316
+ 混
2317
+ 夫
2318
+ 苛
2319
+ 鬚
2320
+ か
2321
+ 杉
2322
+ 游
2323
+ 颪
2324
+ 刊
2325
+ Ⅵ
2326
+ 聡
2327
+ 山
2328
+ 薩
2329
+ 鷹
2330
+ 漫
2331
+ 愈
2332
+ 曹
2333
+ 凌
2334
+ μ
2335
+ 旦
2336
+ 卍
2337
+ 椌
2338
+ 祀
2339
+ 靜
2340
+ 彪
2341
+ 頂
2342
+ 消
2343
+ 贋
2344
+ 荷
2345
+ 煩
2346
+ 倍
2347
+ 琴
2348
+ さ
2349
+ ニ
2350
+ 墓
2351
+ 〇
2352
+ 錦
2353
+ 伜
2354
+ 浦
2355
+ 帥
2356
+ 捗
2357
+ 柑
2358
+ 晏
2359
+ 翳
2360
+ 棄
2361
+ 互
2362
+ 運
2363
+ 椿
2364
+ 峨
2365
+ ぞ
2366
+ 順
2367
+ 憾
2368
+ 欣
2369
+ 晰
2370
+ 保
2371
+ 朗
2372
+ 秘
2373
+ 楜
2374
+ 西
2375
+ 樮
2376
+ 鐘
2377
+ 牟
2378
+ 緒
2379
+ 蟻
2380
+ 鰭
2381
+ 漁
2382
+ ⑫
2383
+ 呉
2384
+ 全
2385
+ 侵
2386
+ ｂ
2387
+ 丈
2388
+ 撲
2389
+ 酢
2390
+ 卑
2391
+ 絡
2392
+ 硫
2393
+ 竹
2394
+ 逸
2395
+ ”
2396
+ 益
2397
+ ―
2398
+ 不
2399
+ 塵
2400
+ 滯
2401
+ 巾
2402
+ 蕨
2403
+ 鰤
2404
+ 亢
2405
+ 版
2406
+ 竪
2407
+ 廟
2408
+ 吠
2409
+ 狸
2410
+ 少
2411
+ 烏
2412
+ 楚
2413
+ 法
2414
+ 亘
2415
+ 魅
2416
+ 布
2417
+ 夷
2418
+ 財
2419
+ 閥
2420
+ 市
2421
+ 挟
2422
+ ⑦
2423
+ 串
2424
+ 剽
2425
+ ？
2426
+ 暮
2427
+ 絃
2428
+ 廃
2429
+ 秤
2430
+ 障
2431
+ 憺
2432
+ 祠
2433
+ 戸
2434
+ 命
2435
+ ら
2436
+ 村
2437
+ 瀨
2438
+ 柔
2439
+ 赤
2440
+ 旗
2441
+ 緖
2442
+ 尾
2443
+ 災
2444
+ 岳
2445
+ 濾
2446
+ 火
2447
+ Ｎ
2448
+ 撒
2449
+ 倒
2450
+ 括
2451
+ 婆
2452
+ 稠
2453
+ 膏
2454
+ 偕
2455
+ 磊
2456
+ 濱
2457
+ 澁
2458
+ 昌
2459
+ ○
2460
+ 鉗
2461
+ 駛
2462
+ 阪
2463
+ 腎
2464
+ 圃
2465
+ 鈴
2466
+ 柚
2467
+ 塒
2468
+ 顆
2469
+ 騙
2470
+ 崎
2471
+ 範
2472
+ 衞
2473
+ 〒
2474
+ 熟
2475
+ 禎
2476
+ 噎
2477
+ 蹉
2478
+ 渉
2479
+ 埼
2480
+ 戒
2481
+ 箱
2482
+ 壺
2483
+ 德
2484
+ 蒐
2485
+ 噌
2486
+ 儒
2487
+ 楴
2488
+ 判
2489
+ 衣
2490
+ 隆
2491
+ 塾
2492
+ う
2493
+ 性
2494
+ 頭
2495
+ 卦
2496
+ 鋭
2497
+ 獅
2498
+ 囂
2499
+ 名
2500
+ 含
2501
+ 郁
2502
+ 届
2503
+ 總
2504
+ 渦
2505
+ 杼
2506
+ 楳
2507
+ 芒
2508
+ 諸
2509
+ 良
2510
+ 萱
2511
+ 傷
2512
+ 麗
2513
+ 做
2514
+ 披
2515
+ 聽
2516
+ 献
2517
+ 腹
2518
+ 必
2519
+ 劃
2520
+ 擽
2521
+ 逼
2522
+ 宗
2523
+ 訶
2524
+ 挽
2525
+ 兌
2526
+ 饗
2527
+ 高
2528
+ 領
2529
+ Δ
2530
+ 出
2531
+ 旨
2532
+ 拒
2533
+ 萸
2534
+ ⑲
2535
+ 盡
2536
+ 攀
2537
+ 附
2538
+ 公
2539
+ 拡
2540
+ 襖
2541
+ 紋
2542
+ 娶
2543
+ 癌
2544
+ 翡
2545
+ 轌
2546
+ 努
2547
+ 断
2548
+ 践
2549
+ 身
2550
+ 舘
2551
+ 霜
2552
+ 庄
2553
+ 怪
2554
+ 壇
2555
+ 芦
2556
+ 辻
2557
+ 飼
2558
+ 纂
2559
+ 嗣
2560
+ 撫
2561
+ Ⅶ
2562
+ 鞄
2563
+ 謝
2564
+ 材
2565
+ 嚢
2566
+ Ｍ
2567
+ 遵
2568
+ 罰
2569
+ ょ
2570
+ 醍
2571
+ 滓
2572
+ 勢
2573
+ 陛
2574
+ 薄
2575
+ て
2576
+ 埓
2577
+ 禮
2578
+ △
2579
+ 肢
2580
+ 掲
2581
+ 嵜
2582
+ 識
2583
+ 匹
2584
+ 幾
2585
+ 麻
2586
+ 琢
2587
+ 奥
2588
+ 砧
2589
+ 價
2590
+ 審
2591
+ 峙
2592
+ Ｄ
2593
+ 嘔
2594
+ 蜚
2595
+ 醉
2596
+ 控
2597
+ 望
2598
+ 歳
2599
+ 姐
2600
+ く
2601
+ 渾
2602
+ 亞
2603
+ 滔
2604
+ 韋
2605
+ 医
2606
+ 僚
2607
+ 觜
2608
+ 矼
2609
+ Ω
2610
+ 懐
2611
+ 婢
2612
+ 對
2613
+ ①
2614
+ 喩
2615
+ 脩
2616
+ 峽
2617
+ 煬
2618
+ ν
2619
+ 猫
2620
+ 掃
2621
+ 攝
2622
+ キ
2623
+ 放
2624
+ 辞
2625
+ 唖
2626
+ 計
2627
+ 蛙
2628
+ 喬
2629
+ 首
2630
+ 薫
2631
+ 蚕
2632
+ 注
2633
+ ！
2634
+ 艶
2635
+ 艫
2636
+ 鉈
2637
+ 紘
2638
+ 殆
2639
+ 糸
2640
+ ヶ
2641
+ 擡
2642
+ 貨
2643
+ 璋
2644
+ 港
2645
+ 罪
2646
+ ぴ
2647
+ 爺
2648
+ 〉
2649
+ 鋤
2650
+ 輾
2651
+ 襄
2652
+ 謨
2653
+ 権
2654
+ 閨
2655
+ ＞
2656
+ 敷
2657
+ 甕
2658
+ 袢
2659
+ 誌
2660
+ 真
2661
+ ⑪
2662
+ 雷
2663
+ 代
2664
+ 梗
2665
+ 魯
2666
+ 釣
2667
+ “
2668
+ 齧
2669
+ 梛
2670
+ 笠
2671
+ 件
2672
+ 副
2673
+ 弟
2674
+ 析
2675
+ 嘆
2676
+ 往
2677
+ 哲
2678
+ 單
2679
+ 八
2680
+ ←
2681
+ 幡
2682
+ 驚
2683
+ 迂
2684
+ 増
2685
+ 坏
2686
+ 別
2687
+ 態
2688
+ 迩
2689
+ 拜
2690
+ 蛮
2691
+ 畸
2692
+ 澎
2693
+ 肘
2694
+ 複
2695
+ 跨
2696
+ 鐸
2697
+ ゃ
2698
+ 戴
2699
+ 桃
2700
+ 嬬
2701
+ 掌
2702
+ 剣
2703
+ 素
2704
+ 為
2705
+ 柤
2706
+ 抛
2707
+ 辺
2708
+ 込
2709
+ 沿
2710
+ 迅
2711
+ 繕
2712
+ 糯
2713
+ 都
2714
+ ↑
2715
+ 奢
2716
+ 鮭
2717
+ 肩
2718
+ 普
2719
+ 賢
2720
+ 敲
2721
+ 砦
2722
+ 8
2723
+ 密
2724
+ 女
2725
+ 氾
2726
+ 蟠
2727
+ 埋
2728
+ 調
2729
+ 力
2730
+ 夏
2731
+ 俳
2732
+ 床
2733
+ 雍
2734
+ ぐ
2735
+ 〃
2736
+ 飲
2737
+ 剩
2738
+ 頼
2739
+ 寄
2740
+ 旬
2741
+ 栓
2742
+ 婚
2743
+ 欝
2744
+ 把
2745
+ 幅
2746
+ 伶
2747
+ 瘤
2748
+ 獄
2749
+ 卓
2750
+ 評
2751
+ 銘
2752
+ 踊
2753
+ 吏
2754
+ 糀
2755
+ ク
2756
+ 悩
2757
+ 鞘
2758
+ 状
2759
+ 鯖
2760
+ こ
2761
+ 憊
2762
+ 綻
2763
+ 斥
2764
+ 閃
2765
+ 寶
2766
+ 垈
2767
+ 筝
2768
+ 綸
2769
+ 碧
2770
+ 是
2771
+ 陶
2772
+ 他
2773
+ 愉
2774
+ 兀
2775
+ 朕
2776
+ 釘
2777
+ 帽
2778
+ 謄
2779
+ 畠
2780
+ 妥
2781
+ 娠
2782
+ 侶
2783
+ 險
2784
+ 会
2785
+ 鴻
2786
+ 黴
2787
+ 遜
2788
+ 体
2789
+ 試
2790
+ 易
2791
+ 帳
2792
+ ぜ
2793
+ 星
2794
+ 珂
2795
+ 鶏
2796
+ 坤
2797
+ 紺
2798
+ °
2799
+ 的
2800
+ 巣
2801
+ 木
2802
+ 捏
2803
+ 隕
2804
+ 廏
2805
+ 簔
2806
+ 艙
2807
+ 勝
2808
+ 娚
2809
+ 玩
2810
+ タ
2811
+ 戟
2812
+ 訛
2813
+ 塩
2814
+ 家
2815
+ 杷
2816
+ 芝
2817
+ 較
2818
+ 楡
2819
+ 舅
2820
+ 閼
2821
+ 浴
2822
+ 漕
2823
+ 囚
2824
+ 孃
2825
+ 渕
2826
+ 翌
2827
+ 灌
2828
+ 灸
2829
+ 遷
2830
+ 皇
2831
+ 島
2832
+ 髭
2833
+ 侮
2834
+ 除
2835
+ 榮
2836
+ 疊
2837
+ 慎
2838
+ 薯
2839
+ 蔀
2840
+ 枅
2841
+ 屯
2842
+ 禽
2843
+ プ
2844
+ 吃
2845
+ 十
2846
+ 忌
2847
+ 也
2848
+ 篶
2849
+ 喀
2850
+ 孔
2851
+ 孜
2852
+ 護
2853
+ 査
2854
+ 渥
2855
+ 擢
2856
+ 謡
2857
+ 燕
2858
+ 歓
2859
+ 塞
2860
+ ア
2861
+ 槍
2862
+ 澄
2863
+ 舷
2864
+ 笏
2865
+ 功
2866
+ 裾
2867
+ 娼
2868
+ 傑
2869
+ 瘍
2870
+ 臺
2871
+ 両
2872
+ 旛
2873
+ 貝
2874
+ 詞
2875
+ 蹂
2876
+ 鄒
2877
+ 榜
2878
+ 場
2879
+ 宵
2880
+ 歩
2881
+ 憂
2882
+ 坑
2883
+ 当
2884
+ 頑
2885
+ 逞
2886
+ す
2887
+ Ⅹ
2888
+ ②
2889
+ 皿
2890
+ 禾
2891
+ Ⅱ
2892
+ 莅
2893
+ 左
2894
+ 慧
2895
+ 蔚
2896
+ 擦
2897
+ 癒
2898
+ 寬
2899
+ ◇
2900
+ 祓
2901
+ 棠
2902
+ 管
2903
+ 販
2904
+ 祝
2905
+ 竇
2906
+ 始
2907
+ 基
2908
+ 一
2909
+ ト
2910
+ 傾
2911
+ 抵
2912
+ 厭
2913
+ 燭
2914
+ 鋒
2915
+ 泙
2916
+ 拵
2917
+ 緞
2918
+ 闔
2919
+ 私
2920
+ 延
2921
+ 紹
2922
+ 圦
2923
+ 蒼
2924
+ 属
2925
+ 枌
2926
+ 応
2927
+ 輪
2928
+ ぁ
2929
+ 諫
2930
+ 挫
2931
+ 膠
2932
+ 皺
2933
+ 亡
2934
+ 棋
2935
+ 幸
2936
+ 僞
2937
+ 脛
2938
+ ゆ
2939
+ 桶
2940
+ 掴
2941
+ 倭
2942
+ 某
2943
+ 頻
2944
+ 岩
2945
+ パ
2946
+ 側
2947
+ 魂
2948
+ 谷
2949
+ 乳
2950
+ 擂
2951
+ 9
2952
+ 冦
2953
+ 峅
2954
+ 圏
2955
+ 驕
2956
+ 胱
2957
+ 咋
2958
+ 穣
2959
+ 酉
2960
+ ル
2961
+ 彫
2962
+ 栖
2963
+ 縣
2964
+ 恐
2965
+ 亟
2966
+ 実
2967
+ 憑
2968
+ 葉
2969
+ 鮨
2970
+ 職
2971
+ 屏
2972
+ 蘭
2973
+ 楓
2974
+ 監
2975
+ 焼
2976
+ 綿
2977
+ 仔
2978
+ 専
2979
+ 廠
2980
+ 気
2981
+ 族
2982
+ 肯
2983
+ 緋
2984
+ 説
2985
+ ヒ
2986
+ サ
2987
+ 曝
2988
+ 酋
2989
+ 用
2990
+ 鱒
2991
+ 香
2992
+ 弯
2993
+ 捌
2994
+ 徘
2995
+ 富
2996
+ 冷
2997
+ 鴛
2998
+ 魁
2999
+ 爛
3000
+ 究
3001
+ 摺
3002
+ 劉
3003
+ 祟
3004
+ 嬰
3005
+ 瀉
3006
+ 挙
3007
+ ｎ
3008
+ 柄
3009
+ 擶
3010
+ ね
3011
+ 聖
3012
+ 兄
3013
+ 泣
3014
+ 瀕
3015
+ 透
3016
+ 逗
3017
+ ば
3018
+ 穢
3019
+ 笹
3020
+ 賓
3021
+ 戦
3022
+ 朴
3023
+ 伴
3024
+ 到
3025
+ 僑
3026
+ 砺
3027
+ 簀
3028
+ 禦
3029
+ 式
3030
+ ん
3031
+ 雀
3032
+ 廿
3033
+ ペ
3034
+ 航
3035
+ み
3036
+ 孝
3037
+ 国
3038
+ 沃
3039
+ 意
3040
+ 甥
3041
+ 詣
3042
+ 楮
3043
+ を
3044
+ 勧
3045
+ 周
3046
+ 軾
3047
+ 狢
3048
+ 紗
3049
+ 啓
3050
+ 胡
3051
+ 臣
3052
+ ‘
3053
+ 詮
3054
+ 檎
3055
+ 晃
3056
+ 鈍
3057
+ 哄
3058
+ 袰
3059
+ 詭
3060
+ 完
3061
+ 享
3062
+ 達
3063
+ 閏
3064
+ 蔓
3065
+ 箇
3066
+ 耘
3067
+ 悄
3068
+ 見
3069
+ 彗
3070
+ 覯
3071
+ 涯
3072
+ 箆
3073
+ 尿
3074
+ 泊
3075
+ 鳥
3076
+ 税
3077
+ ��
3078
+ 滋
3079
+ 壷
3080
+ 称
3081
+ 鬮
3082
+ 嫁
3083
+ 胚
3084
+ 徹
3085
+ 荏
3086
+ 津
3087
+ 濯
3088
+ 墨
3089
+ 数
3090
+ 豆
3091
+ 解
3092
+ 欺
3093
+ 箭
3094
+ 千
3095
+ 賎
3096
+ ／
3097
+ ；
3098
+ 第
3099
+ 呼
3100
+ 凛
3101
+ 組
3102
+ 硯
3103
+ 嫡
3104
+ 割
3105
+ 軍
3106
+ ぇ
3107
+ 持
3108
+ 鮒
3109
+ 企
3110
+ 鋳
3111
+ 滉
3112
+ 埠
3113
+ 刎
3114
+ 燿
3115
+ 景
3116
+ 桓
3117
+ 巡
3118
+ 展
3119
+ 鄙
3120
+ 買
3121
+ 戻
3122
+ 井
3123
+ 撰
3124
+ 毫
3125
+ 菊
3126
+ 禁
3127
+ 畳
3128
+ 嶽
3129
+ 鱗
3130
+ 抗
3131
+ 金
3132
+ 艦
3133
+ 佛
3134
+ 飾
3135
+ 磨
3136
+ 棡
3137
+ 爵
3138
+ 勒
3139
+ 凪
3140
+ 矧
3141
+ 偶
3142
+ 杯
3143
+ 雫
3144
+ 太
3145
+ 奴
3146
+ 於
3147
+ 巻
3148
+ 闖
3149
+ 衡
3150
+ 徳
3151
+ 涅
3152
+ Ｖ
3153
+ 治
3154
+ 洸
3155
+ 訟
3156
+ 葦
3157
+ 悪
3158
+ 唆
3159
+ 溶
3160
+ 慳
3161
+ 扼
3162
+ 笂
3163
+ 縮
3164
+ 鑓
3165
+ 曖
3166
+ 櫃
3167
+ 曜
3168
+ 男
3169
+ 棺
3170
+ 物
3171
+ 施
3172
+ 塘
3173
+ 撚
3174
+ 積
3175
+ ⑳
3176
+ 合
3177
+ 購
3178
+ 鼈
3179
+ 骨
3180
+ 迄
3181
+ 狐
3182
+ 冥
3183
+ 算
3184
+ 飽
3185
+ 湶
3186
+ 狩
3187
+ 活
3188
+ 惨
3189
+ べ
3190
+ 酎
3191
+ 蕎
3192
+ 由
3193
+ 換
3194
+ 霸
3195
+ 偲
3196
+ 靭
3197
+ 耕
3198
+ 倫
3199
+ ボ
3200
+ Ｒ
3201
+ 盧
3202
+ 鐵
3203
+ 営
3204
+ 樺
3205
+ 褄
3206
+ エ
3207
+ 晝
3208
+ 詐
3209
+ 舸
3210
+ 奎
3211
+ 朧
3212
+ 罧
3213
+ 堅
3214
+ 箕
3215
+ 徂
3216
+ 欒
3217
+ 恍
3218
+ 人
3219
+ 啄
3220
+ 油
3221
+ 沁
3222
+ 剥
3223
+ 探
3224
+ 片
3225
+ 万
3226
+ 投
3227
+ 那
3228
+ 芙
3229
+ 娃
3230
+ 碕
3231
+ 些
3232
+ 神
3233
+ 書
3234
+ 俣
3235
+ 邇
3236
+ 刹
3237
+ 柴
3238
+ 本
3239
+ 掎
3240
+ 燧
3241
+ 疲
3242
+ 死
3243
+ 房
3244
+ 惜
3245
+ 地
3246
+ 枷
3247
+ ャ
3248
+ 褶
3249
+ 肱
3250
+ 筰
3251
+ 青
3252
+ ワ
3253
+ 答
3254
+ 斯
3255
+ 昇
3256
+ 明
3257
+ 衲
3258
+ 偏
3259
+ 螂
3260
+ 賑
3261
+ 翹
3262
+ 覧
3263
+ ュ
3264
+ 栢
3265
+ 率
3266
+ 径
3267
+ 惇
3268
+ 点
3269
+ 亦
3270
+ 愍
3271
+ 逮
3272
+ 櫓
3273
+ 克
3274
+ 措
3275
+ 毅
3276
+ 昏
3277
+ 徠
3278
+ 腥
3279
+ 渓
3280
+ 燒
3281
+ 婿
3282
+ 鴬
3283
+ 斐
3284
+ 著
3285
+ 潟
3286
+ 脂
3287
+ 湖
3288
+ 砿
3289
+ 瀾
3290
+ 筧
3291
+ マ
3292
+ 禄
3293
+ 拔
3294
+ 籟
3295
+ 楠
3296
+ 支
3297
+ 哉
3298
+ 僻
3299
+ 燐
3300
+ 黄
3301
+ 贄
3302
+ 晒
3303
+ 曲
3304
+ 遮
3305
+ 砒
3306
+ 乖
3307
+ 覆
3308
+ 又
3309
+ 堰
3310
+ 反
3311
+ 矜
3312
+ 醂
3313
+ 友
3314
+ 新
3315
+ 令
3316
+ 僣
3317
+ 求
3318
+ 賴
3319
+ 醗
3320
+ 丸
3321
+ 姓
3322
+ 愛
3323
+ 裕
3324
+ ゞ
3325
+ 峯
3326
+ 瀞
3327
+ 黑
3328
+ 提
3329
+ 鳳
3330
+ 遭
3331
+ ヴ
3332
+ 喝
3333
+ 菱
3334
+ 趨
3335
+ 診
3336
+ 低
3337
+ 款
3338
+ 枠
3339
+ 城
3340
+ 項
3341
+ 憤
3342
+ 葺
3343
+ 徒
3344
+ 掻
3345
+ 讃
3346
+ 鴫
3347
+ 縞
3348
+ 角
3349
+ 標
3350
+ ▽
3351
+ 斤
3352
+ 粉
3353
+ Ｆ
3354
+ 驀
3355
+ 訣
3356
+ 榑
3357
+ 該
3358
+ 撃
3359
+ 慄
3360
+ 者
3361
+ 菰
3362
+ 鮪
3363
+ 痛
3364
+ 曙
3365
+ 頽
3366
+ 槽
3367
+ 里
3368
+ 住
3369
+ 蛭
3370
+ 促
3371
+ 鍾
3372
+ 携
3373
+ 恥
3374
+ 番
3375
+ 賄
3376
+ 丼
3377
+ 述
3378
+ 最
3379
+ 鍔
3380
+ 弄
3381
+ 朶
3382
+ 芭
3383
+ 堺
3384
+ 肴
3385
+ 珸
3386
+ 船
3387
+ 盈
3388
+ 睡
3389
+ 肇
3390
+ 蔵
3391
+ 袷
3392
+ 儲
3393
+ 河
3394
+ 奪
3395
+ 快
3396
+ 猿
3397
+ 狗
3398
+ 成
3399
+ 戌
3400
+ 短
3401
+ 威
3402
+ 禍
3403
+ 店
3404
+ 漿
3405
+ 画
3406
+ 膵
3407
+ 鼎
3408
+ 畦
3409
+ 夭
3410
+ 萬
3411
+ 籠
3412
+ 濡
3413
+ 讒
3414
+ 雇
3415
+ 鵬
3416
+ 匁
3417
+ 笑
3418
+ 轄
3419
+ 鉢
3420
+ 軟
3421
+ 横
3422
+ 郭
3423
+ √
3424
+ 窒
3425
+ Ｏ
3426
+ 瀦
3427
+ 辱
3428
+ 慰
3429
+ 漏
3430
+ 耳
3431
+ 捧
3432
+ 痙
3433
+ 飄
3434
+ 触
3435
+ ∑
3436
+ 磯
3437
+ つ
3438
+ 亜
3439
+ 宥
3440
+ 脚
3441
+ 剋
3442
+ セ
3443
+ 叙
3444
+ 暁
3445
+ 邂
3446
+ 鳧
3447
+ Ⅸ
3448
+ 﨑
3449
+ 叛
3450
+ 欲
3451
+ 縷
3452
+ Ｗ
3453
+ 紐
3454
+ 劣
3455
+ 仙
3456
+ 彬
3457
+ 椙
3458
+ 面
3459
+ 旧
3460
+ 傲
3461
+ 索
3462
+ 扱
3463
+ 炯
3464
+ 弔
3465
+ 遂
3466
+ 食
3467
+ 感
3468
+ 銃
3469
+ 翠
3470
+ 槐
3471
+ ⑮
3472
+ ≠
3473
+ 温
3474
+ ゲ
3475
+ 榴
3476
+ 懺
3477
+ 〆
3478
+ 虎
3479
+ 士
3480
+ ☆
3481
+ 秡
3482
+ ン
3483
+ 縦
3484
+ 樟
3485
+ 址
3486
+ 搭
3487
+ 賜
3488
+ 鈷
3489
+ 釼
3490
+ 建
3491
+ 伍
3492
+ 鼠
3493
+ 泗
3494
+ 維
3495
+ 漸
3496
+ 否
3497
+ 丞
3498
+ 偵
3499
+ 亨
3500
+ 憚
3501
+ 慇
3502
+ 鐺
3503
+ 葭
3504
+ 七
3505
+ 立
3506
+ 段
3507
+ 殴
3508
+ 桟
3509
+ 話
3510
+ 暗
3511
+ 旅
3512
+ 餌
3513
+ 熊
3514
+ 釈
3515
+ 兔
3516
+ 鎚
3517
+ ζ
3518
+ 粥
3519
+ 禀
3520
+ ホ
3521
+ 甲
3522
+ 汲
3523
+ 乃
3524
+ 擾
3525
+ 岼
3526
+ 舐
3527
+ 茗
3528
+ 諸
3529
+ 鶯
3530
+ 宛
3531
+ 穀
3532
+ 崋
3533
+ 珊
3534
+ 奈
3535
+ 固
3536
+ 幽
3537
+ 蒜
3538
+ 俟
3539
+ ノ
3540
+ 車
3541
+ 吊
3542
+ 赫
3543
+ 厳
3544
+ 國
3545
+ 謎
3546
+ 媚
3547
+ 廂
3548
+ 恕
3549
+ η
3550
+ ♂
3551
+ 鉦
3552
+ 睨
3553
+ 溥
3554
+ 蘂
3555
+ 貿
3556
+ 濫
3557
+ 逍
3558
+ 散
3559
+ 弧
3560
+ 上
3561
+ ご
3562
+ 楕
3563
+ 呻
3564
+ 梳
3565
+ 轆
3566
+ 舎
3567
+ 激
3568
+ 魏
3569
+ 琳
3570
+ 懊
3571
+ 勺
3572
+ 条
3573
+ 鬱
3574
+ 磧
3575
+ 係
3576
+ 拂
3577
+ 界
3578
+ 晶
3579
+ 情
3580
+ 昂
3581
+ 牝
3582
+ 層
3583
+ 黛
3584
+ …
3585
+ 詔
3586
+ 炙
3587
+ 叫
3588
+ 鳩
3589
+ 酒
3590
+ 斎
3591
+ 弊
3592
+ 所
3593
+ 誰
3594
+ 閤
3595
+ 陳
3596
+ 痢
3597
+ 糘
3598
+ 蔦
3599
+ 犢
3600
+ 嚇
3601
+ 菖
3602
+ 馬
3603
+ 祷
3604
+ 洩
3605
+ 免
3606
+ 担
3607
+ 乗
3608
+ 句
3609
+ 翻
3610
+ 轢
3611
+ よ
3612
+ 像
3613
+ 亥
3614
+ ｍ
3615
+ 臨
3616
+ 助
3617
+ 測
3618
+ 竣
3619
+ 差
3620
+ 稔
3621
+ 椴
3622
+ 音
3623
+ ＊
3624
+ 髄
3625
+ 包
3626
+ 頤
3627
+ 蒲
3628
+ 蝋
3629
+ 浬
3630
+ 鴇
3631
+ 轟
3632
+ 唇
3633
+ 悠
3634
+ 棚
3635
+ 蟄
3636
+ 貧
3637
+ 帰
3638
+ 薪
3639
+ 畝
3640
+ 酌
3641
+ 舂
3642
+ 血
3643
+ 堪
3644
+ 耀
3645
+ 暦
3646
+ 教
3647
+ 圈
3648
+ 侭
3649
+ 坡
3650
+ 工
3651
+ 箒
3652
+ 砌
3653
+ 隣
3654
+ 儘
3655
+ ず
3656
+ 浸
3657
+ 實
3658
+ 迪
3659
+ 甜
3660
+ 文
3661
+ 困
3662
+ 誡
3663
+ 福
3664
+ 默
3665
+ 型
3666
+ 償
3667
+ ｊ
3668
+ 鈑
3669
+ 允
3670
+ 弘
3671
+ 貸
3672
+ 塗
3673
+ 廓
3674
+ 秣
3675
+ 久
3676
+ 宙
3677
+ 璧
3678
+ ぅ
3679
+ 糾
3680
+ 擁
3681
+ ，
3682
+ 鯲
3683
+ 荻
3684
+ 裡
3685
+ 覗
3686
+ 畿
3687
+ 穗
3688
+ 毯
3689
+ 俸
3690
+ 橿
3691
+ ：
3692
+ め
3693
+ 雜
3694
+ 育
3695
+ 粟
3696
+ 貫
3697
+ 盂
3698
+ 豚
3699
+ 莵
3700
+ 供
3701
+ 暫
3702
+ 窓
3703
+ 皮
3704
+ 邑
3705
+ 祇
3706
+ 濠
3707
+ 炉
3708
+ 穎
3709
+ ・
3710
+ 搏
3711
+ 若
3712
+ 司
3713
+ 籖
3714
+ 蕊
3715
+ 圷
3716
+ 滝
3717
+ 封
3718
+ 桙
3719
+ ウ
3720
+ 騷
3721
+ 腕
3722
+ 盃
3723
+ 取
3724
+ 鑿
3725
+ 雲
3726
+ 柝
3727
+ 帛
3728
+ 兜
3729
+ 訂
3730
+ 置
3731
+ 砲
3732
+ 静
3733
+ ｉ
3734
+ ぽ
3735
+ 雰
3736
+ 表
3737
+ 憎
3738
+ 曉
3739
+ 諏
3740
+ 訴
3741
+ 顔
3742
+ 切
3743
+ 酔
3744
+ ＆
3745
+ 冬
3746
+ 翼
3747
+ ｃ
3748
+ 視
3749
+ 輩
3750
+ 截
3751
+ 鵤
3752
+ 揶
3753
+ 碍
3754
+ 繁
3755
+ 勾
3756
+ 澪
3757
+ 釀
3758
+ 渇
3759
+ 屍
3760
+ 菓
3761
+ 接
3762
+ 州
3763
+ 疑
3764
+ 悟
3765
+ 季
3766
+ 闇
3767
+ 湯
3768
+ ミ
3769
+ 質
3770
+ 鬘
3771
+ 枝
3772
+ ぼ
3773
+ 躯
3774
+ 矩
3775
+ 祢
3776
+ 白
3777
+ 丹
3778
+ 穴
3779
+ 誼
3780
+ 訪
3781
+ づ
3782
+ 鎗
3783
+ 脇
3784
+ 技
3785
+ ぢ
3786
+ 詠
3787
+ 着
3788
+ 苔
3789
+ 沸
3790
+ 哮
3791
+ 齢
3792
+ 挾
3793
+ 駅
3794
+ 履
3795
+ 府
3796
+ 胸
3797
+ 在
3798
+ 韮
3799
+ 告
3800
+ 発
3801
+ 迫
3802
+ ｜
3803
+ 吾
3804
+ 煽
3805
+ 救
3806
+ 仏
3807
+ ＋
3808
+ 昨
3809
+ 須
3810
+ 搗
3811
+ 藁
3812
+ 梺
3813
+ 京
3814
+ 邁
3815
+ 慙
3816
+ 稜
3817
+ 請
3818
+ 忠
3819
+ 胄
3820
+ 廊
3821
+ 適
3822
+ □
3823
+ 刀
3824
+ 獸
3825
+ 癸
3826
+ 椀
3827
+ 譜
3828
+ 樹
3829
+ 瞼
3830
+ 申
3831
+ 鳫
3832
+ 午
3833
+ 養
3834
+ 鷲
3835
+ 才
3836
+ 渙
3837
+ 蹄
3838
+ 刑
3839
+ 縫
3840
+ 線
3841
+ 壗
3842
+ 從
3843
+ 逆
3844
+ γ
3845
+ 掟
3846
+ 覿
3847
+ 捨
3848
+ 函
3849
+ 馘
3850
+ 未
3851
+ Ｉ
3852
+ 鎮
3853
+ 茨
3854
+ 口
3855
+ 酷
3856
+ ゑ
3857
+ 薬
3858
+ χ
3859
+ 拙
3860
+ 吉
3861
+ ゥ
3862
+ 乞
3863
+ 尚
3864
+ 閾
3865
+ 勁
3866
+ 枇
3867
+ 向
3868
+ ォ
3869
+ 待
3870
+ 枕
3871
+ 掘
3872
+ 残
3873
+ 佐
3874
+ 研
3875
+ 誉
3876
+ 彈
3877
+ 頚
3878
+ 玉
3879
+ 惹
3880
+ 慥
3881
+ 笄
3882
+ 侃
3883
+ 嘴
3884
+ 字
3885
+ 巷
3886
+ 桷
3887
+ 橋
3888
+ ≧
3889
+ 厚
3890
+ 街
3891
+ 乾
3892
+ 擬
3893
+ 澱
3894
+ 圧
3895
+ 跋
3896
+ 姉
3897
+ 歎
3898
+ 癩
3899
+ 掩
3900
+ 莽
3901
+ り
3902
+ 与
3903
+ 澗
3904
+ 彊
3905
+ 舌
3906
+ 賃
3907
+ 糒
3908
+ 頓
3909
+ ι
3910
+ 処
3911
+ 橲
3912
+ 瓜
3913
+ 弛
3914
+ 寧
3915
+ 懸
3916
+ 苦
3917
+ 淳
3918
+ 椚
3919
+ 嗅
3920
+ 硴
3921
+ 錠
3922
+ 禧
3923
+ 損
3924
+ 鴨
3925
+ 麩
3926
+ 胤
3927
+ 爾
3928
+ 埣
3929
+ 嘉
3930
+ υ
3931
+ 堆
3932
+ 縱
3933
+ ぃ
3934
+ 団
3935
+ 學
3936
+ 毒
3937
+ 遣
3938
+ 砂
3939
+ 兒
3940
+ 鮎
3941
+ 翰
3942
+ 創
3943
+ ひ
3944
+ ッ
3945
+ 藺
3946
+ 祿
3947
+ 碩
3948
+ ＄
3949
+ 撼
3950
+ 遽
3951
+ 譴
3952
+ 雹
3953
+ 磁
3954
+ 噸
3955
+ 弥
3956
+ 知
3957
+ 阜
3958
+ 均
3959
+ 璽
3960
+ 妄
3961
+ 域
3962
+ 于
3963
+ 憔
3964
+ 焔
3965
+ 漣
3966
+ 薔
3967
+ 洞
3968
+ 量
3969
+ 庶
3970
+ 甫
3971
+ 傀
3972
+ α
3973
+ 銭
3974
+ 箪
3975
+ 昻
3976
+ 椥
3977
+ 詳
3978
+ 墳
3979
+ 丕
3980
+ 垳
3981
+ 排
3982
+ 関
3983
+ 変
3984
+ 駆
3985
+ 昵
3986
+ 師
3987
+ 雅
3988
+ 芥
3989
+ 棉
3990
+ 瑁
3991
+ 宿
3992
+ 帶
3993
+ 廷
3994
+ 虔
3995
+ 餐
3996
+ 揆
3997
+ 饅
3998
+ 謙
3999
+ コ
4000
+ 神
4001
+ σ
4002
+ 払
4003
+ 鶉
4004
+ っ
4005
+ 灰
4006
+ 薨
4007
+ 綣
4008
+ 元
4009
+ 餉
4010
+ ∧
4011
+ 滌
4012
+ 弾
4013
+ 滑
4014
+ 栴
4015
+ 淀
4016
+ 寓
4017
+ 録
4018
+ 壌
4019
+ テ
4020
+ 務
4021
+ 微
4022
+ 疋
4023
+ 纐
4024
+ 拇
4025
+ 牢
4026
+ 祁
4027
+ 鋸
4028
+ 纜
4029
+ 『
4030
+ イ
4031
+ 軽
4032
+ 射
4033
+ 牧
4034
+ 牙
4035
+ 拿
4036
+ 藝
4037
+ 厄
4038
+ 鬼
4039
+ 特
4040
+ 帖
4041
+ 峰
4042
+ 楷
4043
+ 霄
4044
+ 距
4045
+ 宇
4046
+ 蝦
4047
+ 忽
4048
+ 噛
4049
+ 橡
4050
+ 嶺
4051
+ 誇
4052
+ 巌
4053
+ 号
4054
+ 殿
4055
+ 尼
4056
+ 宮
4057
+ 百
4058
+ 源
4059
+ 稿
4060
+ 遊
4061
+ 筍
4062
+ 鞍
4063
+ ｓ
4064
+ 潜
4065
+ 』
4066
+ 妃
4067
+ 斟
4068
+ 位
4069
+ 乏
4070
+ 檀
4071
+ 拝
4072
+ 裂
4073
+ 祕
4074
+ 爬
4075
+ 嗽
4076
+ 返
4077
+ ム
4078
+ 衿
4079
+ 醒
4080
+ 辰
4081
+ 嘱
4082
+ 遇
4083
+ 尽
4084
+ 謠
4085
+ 駘
4086
+ 阡
4087
+ 憲
4088
+ 猩
4089
+ 鏥
4090
+ 滴
4091
+ 健
4092
+ 現
4093
+ 姦
4094
+ 醤
4095
+ 囃
4096
+ 顯
4097
+ 礼
4098
+ 腔
4099
+ 燃
4100
+ 絆
4101
+ 照
4102
+ 寿
4103
+ 呂
4104
+ 乙
4105
+ 振
4106
+ 矮
4107
+ 突
4108
+ 剤
4109
+ 矯
4110
+ 舳
4111
+ 扶
4112
+ 対
4113
+ 豫
4114
+ 宰
4115
+ ド
4116
+ ヘ
4117
+ ～
4118
+ ξ
4119
+ 兢
4120
+ 榊
4121
+ 補
4122
+ で
4123
+ 尤
4124
+ Ｊ
4125
+ 沙
4126
+ 聯
4127
+ 滲
4128
+ 擲
4129
+ 理
4130
+ 豹
4131
+ 閣
4132
+ 鐙
4133
+ 怒
4134
+ 垪
4135
+ る
4136
+ 絢
4137
+ 獲
4138
+ 莎
4139
+ 蹤
4140
+ 釁
4141
+ 庵
4142
+ 臟
4143
+ 械
4144
+ Ｓ
4145
+ 娘
4146
+ 弓
4147
+ 昆
4148
+ 焉
4149
+ 幣
4150
+ 拗
4151
+ 捷
4152
+ 篩
4153
+ 撮
4154
+ 摘
4155
+ 古
4156
+ 糊
4157
+ 恨
4158
+ 攣
4159
+ 艱
4160
+ 眉
4161
+ 隻
4162
+ 樛
4163
+ 有
4164
+ 播
4165
+ 凰
4166
+ 咤
4167
+ 許
4168
+ 洛
4169
+ 笛
4170
+ 誣
4171
+ 母
4172
+ 舩
4173
+ 讐
4174
+ 降
4175
+ 肌
4176
+ 厨
4177
+ 冗
4178
+ 茉
4179
+ 鰊
4180
+ 鴉
4181
+ 鎖
4182
+ 哺
4183
+ 仰
4184
+ 単
4185
+ 璃
4186
+ 療
4187
+ 川
4188
+ 蛸
4189
+ 椛
4190
+ 狙
4191
+ 害
4192
+ 採
4193
+ 翁
4194
+ 斬
4195
+ 纖
4196
+ φ
4197
+ 燈
4198
+ 瓶
4199
+ 旱
4200
+ 純
4201
+ 且
4202
+ 檗
4203
+ 謳
4204
+ 填
4205
+ 絋
4206
+ 灯
4207
+ 椅
4208
+ 罨
4209
+ 危
4210
+ 需
4211
+ 唱
4212
+ 諛
4213
+ 醸
4214
+ 早
4215
+ 岸
4216
+ 腿
4217
+ 効
4218
+ 卜
4219
+ 速
4220
+ 枢
4221
+ げ
4222
+ 日
4223
+ Ｕ
4224
+ 窩
4225
+ 細
4226
+ 晩
4227
+ 湧
4228
+ 凋
4229
+ 嶋
4230
+ 約
4231
+ 慨
4232
+ 鑑
4233
+ 埒
4234
+ 筈
4235
+ 蝟
4236
+ 梅
4237
+ 印
4238
+ 可
4239
+ 癖
4240
+ 赳
4241
+ 希
4242
+ 竃
4243
+ 鐇
4244
+ 槙
4245
+ 渋
4246
+ 修
4247
+ ．
4248
+ 愼
4249
+ 厩
4250
+ 清
4251
+ 堕
4252
+ 虚
4253
+ 萩
4254
+ 露
4255
+ 勳
4256
+ 呆
4257
+ 蛤
4258
+ 蛾
4259
+ 糠
4260
+ 悌
4261
+ Ｐ
4262
+ 恰
4263
+ 悦
4264
+ 梼
4265
+ ダ
4266
+ 茅
4267
+ 抽
4268
+ 議
4269
+ 尺
4270
+ 博
4271
+ 襟
4272
+ 舒
4273
+ 煉
4274
+ 牽
4275
+ 踪
4276
+ 信
4277
+ 品
4278
+ 民
4279
+ 端
4280
+ 肝
4281
+ 賀
4282
+ 悋
4283
+ 阮
4284
+ ナ
4285
+ Ⅷ
4286
+ 舜
4287
+ 寒
4288
+ 右
4289
+ 惠
4290
+ 腫
4291
+ 哀
4292
+ 帯
4293
+ 筑
4294
+ 喜
4295
+ 敦
4296
+ 梵
4297
+ 韓
4298
+ 籐
4299
+ 缺
4300
+ 氏
4301
+ 暢
4302
+ 岐
4303
+ 訓
4304
+ 魄
4305
+ 殉
4306
+ 盤
4307
+ 霞
4308
+ 幻
4309
+ 台
4310
+ 酪
4311
+ 簇
4312
+ 脈
4313
+ 毘
4314
+ 鎌
4315
+ 脳
4316
+ 躊
4317
+ 逡
4318
+ 奉
4319
+ 揃
4320
+ 入
4321
+ 綾
4322
+ 鋼
4323
+ 泡
4324
+ 猛
4325
+ 恋
4326
+ 膣
4327
+ 辿
4328
+ 笈
4329
+ 度
4330
+ 等
4331
+ 妾
4332
+ 尭
4333
+ 馳
4334
+ 釿
4335
+ 摂
4336
+ 具
4337
+ 侍
4338
+ 瞬
4339
+ 却
4340
+ 配
4341
+ ♀
4342
+ 叺
4343
+ 珠
4344
+ 絹
4345
+ 《
4346
+ 員
4347
+ 汎
4348
+ 昴
4349
+ 築
4350
+ 鼓
4351
+ 然
4352
+ 兇
4353
+ 窺
4354
+ 槇
4355
+ 半
4356
+ 鋺
4357
+ 柞
4358
+ 櫛
4359
+ 英
4360
+ 駈
4361
+ 極
4362
+ 輔
4363
+ 婦
4364
+ ゅ
4365
+ 遍
4366
+ 茸
4367
+ 苣
4368
+ 武
4369
+ ￣
4370
+ ヌ
4371
+ 榎
4372
+ 根
4373
+ 餓
4374
+ ゝ
4375
+ 后
4376
+ 溌
4377
+ 瓢
4378
+ 獪
4379
+ 蟆
4380
+ 慶
4381
+ 要
4382
+ 観
4383
+ 堀
4384
+ 致
4385
+ 荀
4386
+ 鍬
4387
+ 潰
4388
+ 坩
4389
+ 阻
4390
+ 病
4391
+ 鈎
4392
+ 峻
4393
+ 豪
4394
+ 茹
4395
+ 諄
4396
+ 杣
4397
+ 顧
4398
+ ◎
4399
+ 嬌
4400
+ 鶚
4401
+ 回
4402
+ 邪
4403
+ 読
4404
+ 柊
4405
+ 沼
4406
+ 裁
4407
+ 狽
4408
+ 撓
4409
+ 語
4410
+ 即
4411
+ 岨
4412
+ 遏
4413
+ 今
4414
+ 屡
4415
+ 仕
4416
+ 蔭
4417
+ 捺
4418
+ 搖
4419
+ 匠
4420
+ 纒
4421
+ §
4422
+ 討
4423
+ 叶
4424
+ 勿
4425
+ 疎
4426
+ 嗟
4427
+ 裸
4428
+ 厖
4429
+ ′
4430
+ 檜
4431
+ む
4432
+ 察
4433
+ 道
4434
+ 底
4435
+ 愕
4436
+ 呵
4437
+ 汪
4438
+ 菲
4439
+ ﾊ
4440
+ ‐
4441
+ ⇒
models/handwritten-japanese-recognition-0001.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:734967ddeb1037045be348bc6fce8ddf0bfcc601fe2de2d8e8aceeae6239158a
3
+ size 33984614
models/handwritten-japanese-recognition-0001.xml ADDED
@@ -0,0 +1,2252 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <?xml version="1.0"?>
2
+ <net name="torch-jit-export" version="11">
3
+ <layers>
4
+ <layer id="0" name="actual_input" type="Parameter" version="opset1">
5
+ <data shape="1,1,96,2000" element_type="f32" />
6
+ <rt_info>
7
+ <attribute name="old_api_map_element_type" version="0" value="f16" />
8
+ </rt_info>
9
+ <output>
10
+ <port id="0" precision="FP32" names="actual_input">
11
+ <dim>1</dim>
12
+ <dim>1</dim>
13
+ <dim>96</dim>
14
+ <dim>2000</dim>
15
+ <rt_info>
16
+ <attribute name="layout" version="0" layout="[N,C,H,W]" />
17
+ </rt_info>
18
+ </port>
19
+ </output>
20
+ </layer>
21
+ <layer id="1" name="Constant_2751_compressed" type="Const" version="opset1">
22
+ <data element_type="f16" shape="1, 1, 1, 1" offset="0" size="2" />
23
+ <output>
24
+ <port id="0" precision="FP16">
25
+ <dim>1</dim>
26
+ <dim>1</dim>
27
+ <dim>1</dim>
28
+ <dim>1</dim>
29
+ </port>
30
+ </output>
31
+ </layer>
32
+ <layer id="2" name="Constant_2751" type="Convert" version="opset1">
33
+ <data destination_type="f32" />
34
+ <rt_info>
35
+ <attribute name="decompression" version="0" />
36
+ </rt_info>
37
+ <input>
38
+ <port id="0" precision="FP16">
39
+ <dim>1</dim>
40
+ <dim>1</dim>
41
+ <dim>1</dim>
42
+ <dim>1</dim>
43
+ </port>
44
+ </input>
45
+ <output>
46
+ <port id="1" precision="FP32">
47
+ <dim>1</dim>
48
+ <dim>1</dim>
49
+ <dim>1</dim>
50
+ <dim>1</dim>
51
+ </port>
52
+ </output>
53
+ </layer>
54
+ <layer id="3" name="Subtract_351" type="Subtract" version="opset1">
55
+ <data auto_broadcast="numpy" />
56
+ <rt_info>
57
+ <attribute name="preprocessing" version="0" />
58
+ </rt_info>
59
+ <input>
60
+ <port id="0" precision="FP32">
61
+ <dim>1</dim>
62
+ <dim>1</dim>
63
+ <dim>96</dim>
64
+ <dim>2000</dim>
65
+ </port>
66
+ <port id="1" precision="FP32">
67
+ <dim>1</dim>
68
+ <dim>1</dim>
69
+ <dim>1</dim>
70
+ <dim>1</dim>
71
+ </port>
72
+ </input>
73
+ <output>
74
+ <port id="2" precision="FP32">
75
+ <dim>1</dim>
76
+ <dim>1</dim>
77
+ <dim>96</dim>
78
+ <dim>2000</dim>
79
+ </port>
80
+ </output>
81
+ </layer>
82
+ <layer id="4" name="Multiply_2600_compressed" type="Const" version="opset1">
83
+ <data element_type="f16" shape="64, 1, 3, 3" offset="2" size="1152" />
84
+ <rt_info>
85
+ <attribute name="preprocessing" version="0" />
86
+ </rt_info>
87
+ <output>
88
+ <port id="0" precision="FP16">
89
+ <dim>64</dim>
90
+ <dim>1</dim>
91
+ <dim>3</dim>
92
+ <dim>3</dim>
93
+ </port>
94
+ </output>
95
+ </layer>
96
+ <layer id="5" name="Multiply_2600" type="Convert" version="opset1">
97
+ <data destination_type="f32" />
98
+ <rt_info>
99
+ <attribute name="decompression" version="0" />
100
+ <attribute name="preprocessing" version="0" />
101
+ </rt_info>
102
+ <input>
103
+ <port id="0" precision="FP16">
104
+ <dim>64</dim>
105
+ <dim>1</dim>
106
+ <dim>3</dim>
107
+ <dim>3</dim>
108
+ </port>
109
+ </input>
110
+ <output>
111
+ <port id="1" precision="FP32">
112
+ <dim>64</dim>
113
+ <dim>1</dim>
114
+ <dim>3</dim>
115
+ <dim>3</dim>
116
+ </port>
117
+ </output>
118
+ </layer>
119
+ <layer id="6" name="Multiply_2466" type="Convolution" version="opset1">
120
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
121
+ <rt_info>
122
+ <attribute name="preprocessing" version="0" />
123
+ </rt_info>
124
+ <input>
125
+ <port id="0" precision="FP32">
126
+ <dim>1</dim>
127
+ <dim>1</dim>
128
+ <dim>96</dim>
129
+ <dim>2000</dim>
130
+ </port>
131
+ <port id="1" precision="FP32">
132
+ <dim>64</dim>
133
+ <dim>1</dim>
134
+ <dim>3</dim>
135
+ <dim>3</dim>
136
+ </port>
137
+ </input>
138
+ <output>
139
+ <port id="2" precision="FP32">
140
+ <dim>1</dim>
141
+ <dim>64</dim>
142
+ <dim>96</dim>
143
+ <dim>2000</dim>
144
+ </port>
145
+ </output>
146
+ </layer>
147
+ <layer id="7" name="Constant_2471_compressed" type="Const" version="opset1">
148
+ <data element_type="f16" shape="1, 64, 1, 1" offset="1154" size="128" />
149
+ <output>
150
+ <port id="0" precision="FP16">
151
+ <dim>1</dim>
152
+ <dim>64</dim>
153
+ <dim>1</dim>
154
+ <dim>1</dim>
155
+ </port>
156
+ </output>
157
+ </layer>
158
+ <layer id="8" name="Constant_2471" type="Convert" version="opset1">
159
+ <data destination_type="f32" />
160
+ <rt_info>
161
+ <attribute name="decompression" version="0" />
162
+ </rt_info>
163
+ <input>
164
+ <port id="0" precision="FP16">
165
+ <dim>1</dim>
166
+ <dim>64</dim>
167
+ <dim>1</dim>
168
+ <dim>1</dim>
169
+ </port>
170
+ </input>
171
+ <output>
172
+ <port id="1" precision="FP32">
173
+ <dim>1</dim>
174
+ <dim>64</dim>
175
+ <dim>1</dim>
176
+ <dim>1</dim>
177
+ </port>
178
+ </output>
179
+ </layer>
180
+ <layer id="9" name="95" type="Add" version="opset1">
181
+ <data auto_broadcast="numpy" />
182
+ <input>
183
+ <port id="0" precision="FP32">
184
+ <dim>1</dim>
185
+ <dim>64</dim>
186
+ <dim>96</dim>
187
+ <dim>2000</dim>
188
+ </port>
189
+ <port id="1" precision="FP32">
190
+ <dim>1</dim>
191
+ <dim>64</dim>
192
+ <dim>1</dim>
193
+ <dim>1</dim>
194
+ </port>
195
+ </input>
196
+ <output>
197
+ <port id="2" precision="FP32" names="95">
198
+ <dim>1</dim>
199
+ <dim>64</dim>
200
+ <dim>96</dim>
201
+ <dim>2000</dim>
202
+ </port>
203
+ </output>
204
+ </layer>
205
+ <layer id="10" name="96" type="ReLU" version="opset1">
206
+ <input>
207
+ <port id="0" precision="FP32">
208
+ <dim>1</dim>
209
+ <dim>64</dim>
210
+ <dim>96</dim>
211
+ <dim>2000</dim>
212
+ </port>
213
+ </input>
214
+ <output>
215
+ <port id="1" precision="FP32" names="96">
216
+ <dim>1</dim>
217
+ <dim>64</dim>
218
+ <dim>96</dim>
219
+ <dim>2000</dim>
220
+ </port>
221
+ </output>
222
+ </layer>
223
+ <layer id="11" name="Multiply_2606_compressed" type="Const" version="opset1">
224
+ <data element_type="f16" shape="64, 64, 3, 3" offset="1282" size="73728" />
225
+ <output>
226
+ <port id="0" precision="FP16">
227
+ <dim>64</dim>
228
+ <dim>64</dim>
229
+ <dim>3</dim>
230
+ <dim>3</dim>
231
+ </port>
232
+ </output>
233
+ </layer>
234
+ <layer id="12" name="Multiply_2606" type="Convert" version="opset1">
235
+ <data destination_type="f32" />
236
+ <rt_info>
237
+ <attribute name="decompression" version="0" />
238
+ </rt_info>
239
+ <input>
240
+ <port id="0" precision="FP16">
241
+ <dim>64</dim>
242
+ <dim>64</dim>
243
+ <dim>3</dim>
244
+ <dim>3</dim>
245
+ </port>
246
+ </input>
247
+ <output>
248
+ <port id="1" precision="FP32">
249
+ <dim>64</dim>
250
+ <dim>64</dim>
251
+ <dim>3</dim>
252
+ <dim>3</dim>
253
+ </port>
254
+ </output>
255
+ </layer>
256
+ <layer id="13" name="Multiply_2476" type="Convolution" version="opset1">
257
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
258
+ <input>
259
+ <port id="0" precision="FP32">
260
+ <dim>1</dim>
261
+ <dim>64</dim>
262
+ <dim>96</dim>
263
+ <dim>2000</dim>
264
+ </port>
265
+ <port id="1" precision="FP32">
266
+ <dim>64</dim>
267
+ <dim>64</dim>
268
+ <dim>3</dim>
269
+ <dim>3</dim>
270
+ </port>
271
+ </input>
272
+ <output>
273
+ <port id="2" precision="FP32">
274
+ <dim>1</dim>
275
+ <dim>64</dim>
276
+ <dim>96</dim>
277
+ <dim>2000</dim>
278
+ </port>
279
+ </output>
280
+ </layer>
281
+ <layer id="14" name="Constant_2481_compressed" type="Const" version="opset1">
282
+ <data element_type="f16" shape="1, 64, 1, 1" offset="75010" size="128" />
283
+ <output>
284
+ <port id="0" precision="FP16">
285
+ <dim>1</dim>
286
+ <dim>64</dim>
287
+ <dim>1</dim>
288
+ <dim>1</dim>
289
+ </port>
290
+ </output>
291
+ </layer>
292
+ <layer id="15" name="Constant_2481" type="Convert" version="opset1">
293
+ <data destination_type="f32" />
294
+ <rt_info>
295
+ <attribute name="decompression" version="0" />
296
+ </rt_info>
297
+ <input>
298
+ <port id="0" precision="FP16">
299
+ <dim>1</dim>
300
+ <dim>64</dim>
301
+ <dim>1</dim>
302
+ <dim>1</dim>
303
+ </port>
304
+ </input>
305
+ <output>
306
+ <port id="1" precision="FP32">
307
+ <dim>1</dim>
308
+ <dim>64</dim>
309
+ <dim>1</dim>
310
+ <dim>1</dim>
311
+ </port>
312
+ </output>
313
+ </layer>
314
+ <layer id="16" name="98" type="Add" version="opset1">
315
+ <data auto_broadcast="numpy" />
316
+ <input>
317
+ <port id="0" precision="FP32">
318
+ <dim>1</dim>
319
+ <dim>64</dim>
320
+ <dim>96</dim>
321
+ <dim>2000</dim>
322
+ </port>
323
+ <port id="1" precision="FP32">
324
+ <dim>1</dim>
325
+ <dim>64</dim>
326
+ <dim>1</dim>
327
+ <dim>1</dim>
328
+ </port>
329
+ </input>
330
+ <output>
331
+ <port id="2" precision="FP32" names="98">
332
+ <dim>1</dim>
333
+ <dim>64</dim>
334
+ <dim>96</dim>
335
+ <dim>2000</dim>
336
+ </port>
337
+ </output>
338
+ </layer>
339
+ <layer id="17" name="99" type="ReLU" version="opset1">
340
+ <input>
341
+ <port id="0" precision="FP32">
342
+ <dim>1</dim>
343
+ <dim>64</dim>
344
+ <dim>96</dim>
345
+ <dim>2000</dim>
346
+ </port>
347
+ </input>
348
+ <output>
349
+ <port id="1" precision="FP32" names="99">
350
+ <dim>1</dim>
351
+ <dim>64</dim>
352
+ <dim>96</dim>
353
+ <dim>2000</dim>
354
+ </port>
355
+ </output>
356
+ </layer>
357
+ <layer id="18" name="100" type="MaxPool" version="opset8">
358
+ <data strides="2, 2" dilations="1, 1" pads_begin="0, 0" pads_end="0, 0" kernel="2, 2" rounding_type="floor" auto_pad="explicit" index_element_type="i64" axis="0" />
359
+ <input>
360
+ <port id="0" precision="FP32">
361
+ <dim>1</dim>
362
+ <dim>64</dim>
363
+ <dim>96</dim>
364
+ <dim>2000</dim>
365
+ </port>
366
+ </input>
367
+ <output>
368
+ <port id="1" precision="FP32" names="100">
369
+ <dim>1</dim>
370
+ <dim>64</dim>
371
+ <dim>48</dim>
372
+ <dim>1000</dim>
373
+ </port>
374
+ <port id="2" precision="I64">
375
+ <dim>1</dim>
376
+ <dim>64</dim>
377
+ <dim>48</dim>
378
+ <dim>1000</dim>
379
+ </port>
380
+ </output>
381
+ </layer>
382
+ <layer id="19" name="Multiply_2612_compressed" type="Const" version="opset1">
383
+ <data element_type="f16" shape="128, 64, 3, 3" offset="75138" size="147456" />
384
+ <output>
385
+ <port id="0" precision="FP16">
386
+ <dim>128</dim>
387
+ <dim>64</dim>
388
+ <dim>3</dim>
389
+ <dim>3</dim>
390
+ </port>
391
+ </output>
392
+ </layer>
393
+ <layer id="20" name="Multiply_2612" type="Convert" version="opset1">
394
+ <data destination_type="f32" />
395
+ <rt_info>
396
+ <attribute name="decompression" version="0" />
397
+ </rt_info>
398
+ <input>
399
+ <port id="0" precision="FP16">
400
+ <dim>128</dim>
401
+ <dim>64</dim>
402
+ <dim>3</dim>
403
+ <dim>3</dim>
404
+ </port>
405
+ </input>
406
+ <output>
407
+ <port id="1" precision="FP32">
408
+ <dim>128</dim>
409
+ <dim>64</dim>
410
+ <dim>3</dim>
411
+ <dim>3</dim>
412
+ </port>
413
+ </output>
414
+ </layer>
415
+ <layer id="21" name="Multiply_2486" type="Convolution" version="opset1">
416
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
417
+ <input>
418
+ <port id="0" precision="FP32">
419
+ <dim>1</dim>
420
+ <dim>64</dim>
421
+ <dim>48</dim>
422
+ <dim>1000</dim>
423
+ </port>
424
+ <port id="1" precision="FP32">
425
+ <dim>128</dim>
426
+ <dim>64</dim>
427
+ <dim>3</dim>
428
+ <dim>3</dim>
429
+ </port>
430
+ </input>
431
+ <output>
432
+ <port id="2" precision="FP32">
433
+ <dim>1</dim>
434
+ <dim>128</dim>
435
+ <dim>48</dim>
436
+ <dim>1000</dim>
437
+ </port>
438
+ </output>
439
+ </layer>
440
+ <layer id="22" name="Constant_2491_compressed" type="Const" version="opset1">
441
+ <data element_type="f16" shape="1, 128, 1, 1" offset="222594" size="256" />
442
+ <output>
443
+ <port id="0" precision="FP16">
444
+ <dim>1</dim>
445
+ <dim>128</dim>
446
+ <dim>1</dim>
447
+ <dim>1</dim>
448
+ </port>
449
+ </output>
450
+ </layer>
451
+ <layer id="23" name="Constant_2491" type="Convert" version="opset1">
452
+ <data destination_type="f32" />
453
+ <rt_info>
454
+ <attribute name="decompression" version="0" />
455
+ </rt_info>
456
+ <input>
457
+ <port id="0" precision="FP16">
458
+ <dim>1</dim>
459
+ <dim>128</dim>
460
+ <dim>1</dim>
461
+ <dim>1</dim>
462
+ </port>
463
+ </input>
464
+ <output>
465
+ <port id="1" precision="FP32">
466
+ <dim>1</dim>
467
+ <dim>128</dim>
468
+ <dim>1</dim>
469
+ <dim>1</dim>
470
+ </port>
471
+ </output>
472
+ </layer>
473
+ <layer id="24" name="102" type="Add" version="opset1">
474
+ <data auto_broadcast="numpy" />
475
+ <input>
476
+ <port id="0" precision="FP32">
477
+ <dim>1</dim>
478
+ <dim>128</dim>
479
+ <dim>48</dim>
480
+ <dim>1000</dim>
481
+ </port>
482
+ <port id="1" precision="FP32">
483
+ <dim>1</dim>
484
+ <dim>128</dim>
485
+ <dim>1</dim>
486
+ <dim>1</dim>
487
+ </port>
488
+ </input>
489
+ <output>
490
+ <port id="2" precision="FP32" names="102">
491
+ <dim>1</dim>
492
+ <dim>128</dim>
493
+ <dim>48</dim>
494
+ <dim>1000</dim>
495
+ </port>
496
+ </output>
497
+ </layer>
498
+ <layer id="25" name="103" type="ReLU" version="opset1">
499
+ <input>
500
+ <port id="0" precision="FP32">
501
+ <dim>1</dim>
502
+ <dim>128</dim>
503
+ <dim>48</dim>
504
+ <dim>1000</dim>
505
+ </port>
506
+ </input>
507
+ <output>
508
+ <port id="1" precision="FP32" names="103">
509
+ <dim>1</dim>
510
+ <dim>128</dim>
511
+ <dim>48</dim>
512
+ <dim>1000</dim>
513
+ </port>
514
+ </output>
515
+ </layer>
516
+ <layer id="26" name="Multiply_2618_compressed" type="Const" version="opset1">
517
+ <data element_type="f16" shape="128, 128, 3, 3" offset="222850" size="294912" />
518
+ <output>
519
+ <port id="0" precision="FP16">
520
+ <dim>128</dim>
521
+ <dim>128</dim>
522
+ <dim>3</dim>
523
+ <dim>3</dim>
524
+ </port>
525
+ </output>
526
+ </layer>
527
+ <layer id="27" name="Multiply_2618" type="Convert" version="opset1">
528
+ <data destination_type="f32" />
529
+ <rt_info>
530
+ <attribute name="decompression" version="0" />
531
+ </rt_info>
532
+ <input>
533
+ <port id="0" precision="FP16">
534
+ <dim>128</dim>
535
+ <dim>128</dim>
536
+ <dim>3</dim>
537
+ <dim>3</dim>
538
+ </port>
539
+ </input>
540
+ <output>
541
+ <port id="1" precision="FP32">
542
+ <dim>128</dim>
543
+ <dim>128</dim>
544
+ <dim>3</dim>
545
+ <dim>3</dim>
546
+ </port>
547
+ </output>
548
+ </layer>
549
+ <layer id="28" name="Multiply_2496" type="Convolution" version="opset1">
550
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
551
+ <input>
552
+ <port id="0" precision="FP32">
553
+ <dim>1</dim>
554
+ <dim>128</dim>
555
+ <dim>48</dim>
556
+ <dim>1000</dim>
557
+ </port>
558
+ <port id="1" precision="FP32">
559
+ <dim>128</dim>
560
+ <dim>128</dim>
561
+ <dim>3</dim>
562
+ <dim>3</dim>
563
+ </port>
564
+ </input>
565
+ <output>
566
+ <port id="2" precision="FP32">
567
+ <dim>1</dim>
568
+ <dim>128</dim>
569
+ <dim>48</dim>
570
+ <dim>1000</dim>
571
+ </port>
572
+ </output>
573
+ </layer>
574
+ <layer id="29" name="Constant_2501_compressed" type="Const" version="opset1">
575
+ <data element_type="f16" shape="1, 128, 1, 1" offset="517762" size="256" />
576
+ <output>
577
+ <port id="0" precision="FP16">
578
+ <dim>1</dim>
579
+ <dim>128</dim>
580
+ <dim>1</dim>
581
+ <dim>1</dim>
582
+ </port>
583
+ </output>
584
+ </layer>
585
+ <layer id="30" name="Constant_2501" type="Convert" version="opset1">
586
+ <data destination_type="f32" />
587
+ <rt_info>
588
+ <attribute name="decompression" version="0" />
589
+ </rt_info>
590
+ <input>
591
+ <port id="0" precision="FP16">
592
+ <dim>1</dim>
593
+ <dim>128</dim>
594
+ <dim>1</dim>
595
+ <dim>1</dim>
596
+ </port>
597
+ </input>
598
+ <output>
599
+ <port id="1" precision="FP32">
600
+ <dim>1</dim>
601
+ <dim>128</dim>
602
+ <dim>1</dim>
603
+ <dim>1</dim>
604
+ </port>
605
+ </output>
606
+ </layer>
607
+ <layer id="31" name="105" type="Add" version="opset1">
608
+ <data auto_broadcast="numpy" />
609
+ <input>
610
+ <port id="0" precision="FP32">
611
+ <dim>1</dim>
612
+ <dim>128</dim>
613
+ <dim>48</dim>
614
+ <dim>1000</dim>
615
+ </port>
616
+ <port id="1" precision="FP32">
617
+ <dim>1</dim>
618
+ <dim>128</dim>
619
+ <dim>1</dim>
620
+ <dim>1</dim>
621
+ </port>
622
+ </input>
623
+ <output>
624
+ <port id="2" precision="FP32" names="105">
625
+ <dim>1</dim>
626
+ <dim>128</dim>
627
+ <dim>48</dim>
628
+ <dim>1000</dim>
629
+ </port>
630
+ </output>
631
+ </layer>
632
+ <layer id="32" name="106" type="ReLU" version="opset1">
633
+ <input>
634
+ <port id="0" precision="FP32">
635
+ <dim>1</dim>
636
+ <dim>128</dim>
637
+ <dim>48</dim>
638
+ <dim>1000</dim>
639
+ </port>
640
+ </input>
641
+ <output>
642
+ <port id="1" precision="FP32" names="106">
643
+ <dim>1</dim>
644
+ <dim>128</dim>
645
+ <dim>48</dim>
646
+ <dim>1000</dim>
647
+ </port>
648
+ </output>
649
+ </layer>
650
+ <layer id="33" name="107" type="MaxPool" version="opset8">
651
+ <data strides="2, 2" dilations="1, 1" pads_begin="0, 0" pads_end="0, 0" kernel="2, 2" rounding_type="floor" auto_pad="explicit" index_element_type="i64" axis="0" />
652
+ <input>
653
+ <port id="0" precision="FP32">
654
+ <dim>1</dim>
655
+ <dim>128</dim>
656
+ <dim>48</dim>
657
+ <dim>1000</dim>
658
+ </port>
659
+ </input>
660
+ <output>
661
+ <port id="1" precision="FP32" names="107">
662
+ <dim>1</dim>
663
+ <dim>128</dim>
664
+ <dim>24</dim>
665
+ <dim>500</dim>
666
+ </port>
667
+ <port id="2" precision="I64">
668
+ <dim>1</dim>
669
+ <dim>128</dim>
670
+ <dim>24</dim>
671
+ <dim>500</dim>
672
+ </port>
673
+ </output>
674
+ </layer>
675
+ <layer id="34" name="Multiply_2624_compressed" type="Const" version="opset1">
676
+ <data element_type="f16" shape="256, 128, 3, 3" offset="518018" size="589824" />
677
+ <output>
678
+ <port id="0" precision="FP16">
679
+ <dim>256</dim>
680
+ <dim>128</dim>
681
+ <dim>3</dim>
682
+ <dim>3</dim>
683
+ </port>
684
+ </output>
685
+ </layer>
686
+ <layer id="35" name="Multiply_2624" type="Convert" version="opset1">
687
+ <data destination_type="f32" />
688
+ <rt_info>
689
+ <attribute name="decompression" version="0" />
690
+ </rt_info>
691
+ <input>
692
+ <port id="0" precision="FP16">
693
+ <dim>256</dim>
694
+ <dim>128</dim>
695
+ <dim>3</dim>
696
+ <dim>3</dim>
697
+ </port>
698
+ </input>
699
+ <output>
700
+ <port id="1" precision="FP32">
701
+ <dim>256</dim>
702
+ <dim>128</dim>
703
+ <dim>3</dim>
704
+ <dim>3</dim>
705
+ </port>
706
+ </output>
707
+ </layer>
708
+ <layer id="36" name="Multiply_2506" type="Convolution" version="opset1">
709
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
710
+ <input>
711
+ <port id="0" precision="FP32">
712
+ <dim>1</dim>
713
+ <dim>128</dim>
714
+ <dim>24</dim>
715
+ <dim>500</dim>
716
+ </port>
717
+ <port id="1" precision="FP32">
718
+ <dim>256</dim>
719
+ <dim>128</dim>
720
+ <dim>3</dim>
721
+ <dim>3</dim>
722
+ </port>
723
+ </input>
724
+ <output>
725
+ <port id="2" precision="FP32">
726
+ <dim>1</dim>
727
+ <dim>256</dim>
728
+ <dim>24</dim>
729
+ <dim>500</dim>
730
+ </port>
731
+ </output>
732
+ </layer>
733
+ <layer id="37" name="Constant_2511_compressed" type="Const" version="opset1">
734
+ <data element_type="f16" shape="1, 256, 1, 1" offset="1107842" size="512" />
735
+ <output>
736
+ <port id="0" precision="FP16">
737
+ <dim>1</dim>
738
+ <dim>256</dim>
739
+ <dim>1</dim>
740
+ <dim>1</dim>
741
+ </port>
742
+ </output>
743
+ </layer>
744
+ <layer id="38" name="Constant_2511" type="Convert" version="opset1">
745
+ <data destination_type="f32" />
746
+ <rt_info>
747
+ <attribute name="decompression" version="0" />
748
+ </rt_info>
749
+ <input>
750
+ <port id="0" precision="FP16">
751
+ <dim>1</dim>
752
+ <dim>256</dim>
753
+ <dim>1</dim>
754
+ <dim>1</dim>
755
+ </port>
756
+ </input>
757
+ <output>
758
+ <port id="1" precision="FP32">
759
+ <dim>1</dim>
760
+ <dim>256</dim>
761
+ <dim>1</dim>
762
+ <dim>1</dim>
763
+ </port>
764
+ </output>
765
+ </layer>
766
+ <layer id="39" name="109" type="Add" version="opset1">
767
+ <data auto_broadcast="numpy" />
768
+ <input>
769
+ <port id="0" precision="FP32">
770
+ <dim>1</dim>
771
+ <dim>256</dim>
772
+ <dim>24</dim>
773
+ <dim>500</dim>
774
+ </port>
775
+ <port id="1" precision="FP32">
776
+ <dim>1</dim>
777
+ <dim>256</dim>
778
+ <dim>1</dim>
779
+ <dim>1</dim>
780
+ </port>
781
+ </input>
782
+ <output>
783
+ <port id="2" precision="FP32" names="109">
784
+ <dim>1</dim>
785
+ <dim>256</dim>
786
+ <dim>24</dim>
787
+ <dim>500</dim>
788
+ </port>
789
+ </output>
790
+ </layer>
791
+ <layer id="40" name="110" type="ReLU" version="opset1">
792
+ <input>
793
+ <port id="0" precision="FP32">
794
+ <dim>1</dim>
795
+ <dim>256</dim>
796
+ <dim>24</dim>
797
+ <dim>500</dim>
798
+ </port>
799
+ </input>
800
+ <output>
801
+ <port id="1" precision="FP32" names="110">
802
+ <dim>1</dim>
803
+ <dim>256</dim>
804
+ <dim>24</dim>
805
+ <dim>500</dim>
806
+ </port>
807
+ </output>
808
+ </layer>
809
+ <layer id="41" name="Multiply_2630_compressed" type="Const" version="opset1">
810
+ <data element_type="f16" shape="256, 256, 3, 3" offset="1108354" size="1179648" />
811
+ <output>
812
+ <port id="0" precision="FP16">
813
+ <dim>256</dim>
814
+ <dim>256</dim>
815
+ <dim>3</dim>
816
+ <dim>3</dim>
817
+ </port>
818
+ </output>
819
+ </layer>
820
+ <layer id="42" name="Multiply_2630" type="Convert" version="opset1">
821
+ <data destination_type="f32" />
822
+ <rt_info>
823
+ <attribute name="decompression" version="0" />
824
+ </rt_info>
825
+ <input>
826
+ <port id="0" precision="FP16">
827
+ <dim>256</dim>
828
+ <dim>256</dim>
829
+ <dim>3</dim>
830
+ <dim>3</dim>
831
+ </port>
832
+ </input>
833
+ <output>
834
+ <port id="1" precision="FP32">
835
+ <dim>256</dim>
836
+ <dim>256</dim>
837
+ <dim>3</dim>
838
+ <dim>3</dim>
839
+ </port>
840
+ </output>
841
+ </layer>
842
+ <layer id="43" name="Multiply_2516" type="Convolution" version="opset1">
843
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
844
+ <input>
845
+ <port id="0" precision="FP32">
846
+ <dim>1</dim>
847
+ <dim>256</dim>
848
+ <dim>24</dim>
849
+ <dim>500</dim>
850
+ </port>
851
+ <port id="1" precision="FP32">
852
+ <dim>256</dim>
853
+ <dim>256</dim>
854
+ <dim>3</dim>
855
+ <dim>3</dim>
856
+ </port>
857
+ </input>
858
+ <output>
859
+ <port id="2" precision="FP32">
860
+ <dim>1</dim>
861
+ <dim>256</dim>
862
+ <dim>24</dim>
863
+ <dim>500</dim>
864
+ </port>
865
+ </output>
866
+ </layer>
867
+ <layer id="44" name="Constant_2521_compressed" type="Const" version="opset1">
868
+ <data element_type="f16" shape="1, 256, 1, 1" offset="2288002" size="512" />
869
+ <output>
870
+ <port id="0" precision="FP16">
871
+ <dim>1</dim>
872
+ <dim>256</dim>
873
+ <dim>1</dim>
874
+ <dim>1</dim>
875
+ </port>
876
+ </output>
877
+ </layer>
878
+ <layer id="45" name="Constant_2521" type="Convert" version="opset1">
879
+ <data destination_type="f32" />
880
+ <rt_info>
881
+ <attribute name="decompression" version="0" />
882
+ </rt_info>
883
+ <input>
884
+ <port id="0" precision="FP16">
885
+ <dim>1</dim>
886
+ <dim>256</dim>
887
+ <dim>1</dim>
888
+ <dim>1</dim>
889
+ </port>
890
+ </input>
891
+ <output>
892
+ <port id="1" precision="FP32">
893
+ <dim>1</dim>
894
+ <dim>256</dim>
895
+ <dim>1</dim>
896
+ <dim>1</dim>
897
+ </port>
898
+ </output>
899
+ </layer>
900
+ <layer id="46" name="112" type="Add" version="opset1">
901
+ <data auto_broadcast="numpy" />
902
+ <input>
903
+ <port id="0" precision="FP32">
904
+ <dim>1</dim>
905
+ <dim>256</dim>
906
+ <dim>24</dim>
907
+ <dim>500</dim>
908
+ </port>
909
+ <port id="1" precision="FP32">
910
+ <dim>1</dim>
911
+ <dim>256</dim>
912
+ <dim>1</dim>
913
+ <dim>1</dim>
914
+ </port>
915
+ </input>
916
+ <output>
917
+ <port id="2" precision="FP32" names="112">
918
+ <dim>1</dim>
919
+ <dim>256</dim>
920
+ <dim>24</dim>
921
+ <dim>500</dim>
922
+ </port>
923
+ </output>
924
+ </layer>
925
+ <layer id="47" name="113" type="ReLU" version="opset1">
926
+ <input>
927
+ <port id="0" precision="FP32">
928
+ <dim>1</dim>
929
+ <dim>256</dim>
930
+ <dim>24</dim>
931
+ <dim>500</dim>
932
+ </port>
933
+ </input>
934
+ <output>
935
+ <port id="1" precision="FP32" names="113">
936
+ <dim>1</dim>
937
+ <dim>256</dim>
938
+ <dim>24</dim>
939
+ <dim>500</dim>
940
+ </port>
941
+ </output>
942
+ </layer>
943
+ <layer id="48" name="Multiply_2636_compressed" type="Const" version="opset1">
944
+ <data element_type="f16" shape="256, 256, 3, 3" offset="2288514" size="1179648" />
945
+ <output>
946
+ <port id="0" precision="FP16">
947
+ <dim>256</dim>
948
+ <dim>256</dim>
949
+ <dim>3</dim>
950
+ <dim>3</dim>
951
+ </port>
952
+ </output>
953
+ </layer>
954
+ <layer id="49" name="Multiply_2636" type="Convert" version="opset1">
955
+ <data destination_type="f32" />
956
+ <rt_info>
957
+ <attribute name="decompression" version="0" />
958
+ </rt_info>
959
+ <input>
960
+ <port id="0" precision="FP16">
961
+ <dim>256</dim>
962
+ <dim>256</dim>
963
+ <dim>3</dim>
964
+ <dim>3</dim>
965
+ </port>
966
+ </input>
967
+ <output>
968
+ <port id="1" precision="FP32">
969
+ <dim>256</dim>
970
+ <dim>256</dim>
971
+ <dim>3</dim>
972
+ <dim>3</dim>
973
+ </port>
974
+ </output>
975
+ </layer>
976
+ <layer id="50" name="Multiply_2526" type="Convolution" version="opset1">
977
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
978
+ <input>
979
+ <port id="0" precision="FP32">
980
+ <dim>1</dim>
981
+ <dim>256</dim>
982
+ <dim>24</dim>
983
+ <dim>500</dim>
984
+ </port>
985
+ <port id="1" precision="FP32">
986
+ <dim>256</dim>
987
+ <dim>256</dim>
988
+ <dim>3</dim>
989
+ <dim>3</dim>
990
+ </port>
991
+ </input>
992
+ <output>
993
+ <port id="2" precision="FP32">
994
+ <dim>1</dim>
995
+ <dim>256</dim>
996
+ <dim>24</dim>
997
+ <dim>500</dim>
998
+ </port>
999
+ </output>
1000
+ </layer>
1001
+ <layer id="51" name="Constant_2531_compressed" type="Const" version="opset1">
1002
+ <data element_type="f16" shape="1, 256, 1, 1" offset="3468162" size="512" />
1003
+ <output>
1004
+ <port id="0" precision="FP16">
1005
+ <dim>1</dim>
1006
+ <dim>256</dim>
1007
+ <dim>1</dim>
1008
+ <dim>1</dim>
1009
+ </port>
1010
+ </output>
1011
+ </layer>
1012
+ <layer id="52" name="Constant_2531" type="Convert" version="opset1">
1013
+ <data destination_type="f32" />
1014
+ <rt_info>
1015
+ <attribute name="decompression" version="0" />
1016
+ </rt_info>
1017
+ <input>
1018
+ <port id="0" precision="FP16">
1019
+ <dim>1</dim>
1020
+ <dim>256</dim>
1021
+ <dim>1</dim>
1022
+ <dim>1</dim>
1023
+ </port>
1024
+ </input>
1025
+ <output>
1026
+ <port id="1" precision="FP32">
1027
+ <dim>1</dim>
1028
+ <dim>256</dim>
1029
+ <dim>1</dim>
1030
+ <dim>1</dim>
1031
+ </port>
1032
+ </output>
1033
+ </layer>
1034
+ <layer id="53" name="115" type="Add" version="opset1">
1035
+ <data auto_broadcast="numpy" />
1036
+ <input>
1037
+ <port id="0" precision="FP32">
1038
+ <dim>1</dim>
1039
+ <dim>256</dim>
1040
+ <dim>24</dim>
1041
+ <dim>500</dim>
1042
+ </port>
1043
+ <port id="1" precision="FP32">
1044
+ <dim>1</dim>
1045
+ <dim>256</dim>
1046
+ <dim>1</dim>
1047
+ <dim>1</dim>
1048
+ </port>
1049
+ </input>
1050
+ <output>
1051
+ <port id="2" precision="FP32" names="115">
1052
+ <dim>1</dim>
1053
+ <dim>256</dim>
1054
+ <dim>24</dim>
1055
+ <dim>500</dim>
1056
+ </port>
1057
+ </output>
1058
+ </layer>
1059
+ <layer id="54" name="116" type="ReLU" version="opset1">
1060
+ <input>
1061
+ <port id="0" precision="FP32">
1062
+ <dim>1</dim>
1063
+ <dim>256</dim>
1064
+ <dim>24</dim>
1065
+ <dim>500</dim>
1066
+ </port>
1067
+ </input>
1068
+ <output>
1069
+ <port id="1" precision="FP32" names="116">
1070
+ <dim>1</dim>
1071
+ <dim>256</dim>
1072
+ <dim>24</dim>
1073
+ <dim>500</dim>
1074
+ </port>
1075
+ </output>
1076
+ </layer>
1077
+ <layer id="55" name="118" type="MaxPool" version="opset8">
1078
+ <data strides="2, 2" dilations="1, 1" pads_begin="0, 0" pads_end="0, 0" kernel="2, 2" rounding_type="floor" auto_pad="explicit" index_element_type="i64" axis="0" />
1079
+ <input>
1080
+ <port id="0" precision="FP32">
1081
+ <dim>1</dim>
1082
+ <dim>256</dim>
1083
+ <dim>24</dim>
1084
+ <dim>500</dim>
1085
+ </port>
1086
+ </input>
1087
+ <output>
1088
+ <port id="1" precision="FP32" names="118">
1089
+ <dim>1</dim>
1090
+ <dim>256</dim>
1091
+ <dim>12</dim>
1092
+ <dim>250</dim>
1093
+ </port>
1094
+ <port id="2" precision="I64">
1095
+ <dim>1</dim>
1096
+ <dim>256</dim>
1097
+ <dim>12</dim>
1098
+ <dim>250</dim>
1099
+ </port>
1100
+ </output>
1101
+ </layer>
1102
+ <layer id="56" name="Multiply_2642_compressed" type="Const" version="opset1">
1103
+ <data element_type="f16" shape="512, 256, 3, 3" offset="3468674" size="2359296" />
1104
+ <output>
1105
+ <port id="0" precision="FP16">
1106
+ <dim>512</dim>
1107
+ <dim>256</dim>
1108
+ <dim>3</dim>
1109
+ <dim>3</dim>
1110
+ </port>
1111
+ </output>
1112
+ </layer>
1113
+ <layer id="57" name="Multiply_2642" type="Convert" version="opset1">
1114
+ <data destination_type="f32" />
1115
+ <rt_info>
1116
+ <attribute name="decompression" version="0" />
1117
+ </rt_info>
1118
+ <input>
1119
+ <port id="0" precision="FP16">
1120
+ <dim>512</dim>
1121
+ <dim>256</dim>
1122
+ <dim>3</dim>
1123
+ <dim>3</dim>
1124
+ </port>
1125
+ </input>
1126
+ <output>
1127
+ <port id="1" precision="FP32">
1128
+ <dim>512</dim>
1129
+ <dim>256</dim>
1130
+ <dim>3</dim>
1131
+ <dim>3</dim>
1132
+ </port>
1133
+ </output>
1134
+ </layer>
1135
+ <layer id="58" name="Multiply_2536" type="Convolution" version="opset1">
1136
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
1137
+ <input>
1138
+ <port id="0" precision="FP32">
1139
+ <dim>1</dim>
1140
+ <dim>256</dim>
1141
+ <dim>12</dim>
1142
+ <dim>250</dim>
1143
+ </port>
1144
+ <port id="1" precision="FP32">
1145
+ <dim>512</dim>
1146
+ <dim>256</dim>
1147
+ <dim>3</dim>
1148
+ <dim>3</dim>
1149
+ </port>
1150
+ </input>
1151
+ <output>
1152
+ <port id="2" precision="FP32">
1153
+ <dim>1</dim>
1154
+ <dim>512</dim>
1155
+ <dim>12</dim>
1156
+ <dim>250</dim>
1157
+ </port>
1158
+ </output>
1159
+ </layer>
1160
+ <layer id="59" name="Constant_2541_compressed" type="Const" version="opset1">
1161
+ <data element_type="f16" shape="1, 512, 1, 1" offset="5827970" size="1024" />
1162
+ <output>
1163
+ <port id="0" precision="FP16">
1164
+ <dim>1</dim>
1165
+ <dim>512</dim>
1166
+ <dim>1</dim>
1167
+ <dim>1</dim>
1168
+ </port>
1169
+ </output>
1170
+ </layer>
1171
+ <layer id="60" name="Constant_2541" type="Convert" version="opset1">
1172
+ <data destination_type="f32" />
1173
+ <rt_info>
1174
+ <attribute name="decompression" version="0" />
1175
+ </rt_info>
1176
+ <input>
1177
+ <port id="0" precision="FP16">
1178
+ <dim>1</dim>
1179
+ <dim>512</dim>
1180
+ <dim>1</dim>
1181
+ <dim>1</dim>
1182
+ </port>
1183
+ </input>
1184
+ <output>
1185
+ <port id="1" precision="FP32">
1186
+ <dim>1</dim>
1187
+ <dim>512</dim>
1188
+ <dim>1</dim>
1189
+ <dim>1</dim>
1190
+ </port>
1191
+ </output>
1192
+ </layer>
1193
+ <layer id="61" name="121" type="Add" version="opset1">
1194
+ <data auto_broadcast="numpy" />
1195
+ <input>
1196
+ <port id="0" precision="FP32">
1197
+ <dim>1</dim>
1198
+ <dim>512</dim>
1199
+ <dim>12</dim>
1200
+ <dim>250</dim>
1201
+ </port>
1202
+ <port id="1" precision="FP32">
1203
+ <dim>1</dim>
1204
+ <dim>512</dim>
1205
+ <dim>1</dim>
1206
+ <dim>1</dim>
1207
+ </port>
1208
+ </input>
1209
+ <output>
1210
+ <port id="2" precision="FP32" names="121">
1211
+ <dim>1</dim>
1212
+ <dim>512</dim>
1213
+ <dim>12</dim>
1214
+ <dim>250</dim>
1215
+ </port>
1216
+ </output>
1217
+ </layer>
1218
+ <layer id="62" name="122" type="ReLU" version="opset1">
1219
+ <input>
1220
+ <port id="0" precision="FP32">
1221
+ <dim>1</dim>
1222
+ <dim>512</dim>
1223
+ <dim>12</dim>
1224
+ <dim>250</dim>
1225
+ </port>
1226
+ </input>
1227
+ <output>
1228
+ <port id="1" precision="FP32" names="122">
1229
+ <dim>1</dim>
1230
+ <dim>512</dim>
1231
+ <dim>12</dim>
1232
+ <dim>250</dim>
1233
+ </port>
1234
+ </output>
1235
+ </layer>
1236
+ <layer id="63" name="Multiply_2648_compressed" type="Const" version="opset1">
1237
+ <data element_type="f16" shape="512, 512, 3, 3" offset="5828994" size="4718592" />
1238
+ <output>
1239
+ <port id="0" precision="FP16">
1240
+ <dim>512</dim>
1241
+ <dim>512</dim>
1242
+ <dim>3</dim>
1243
+ <dim>3</dim>
1244
+ </port>
1245
+ </output>
1246
+ </layer>
1247
+ <layer id="64" name="Multiply_2648" type="Convert" version="opset1">
1248
+ <data destination_type="f32" />
1249
+ <rt_info>
1250
+ <attribute name="decompression" version="0" />
1251
+ </rt_info>
1252
+ <input>
1253
+ <port id="0" precision="FP16">
1254
+ <dim>512</dim>
1255
+ <dim>512</dim>
1256
+ <dim>3</dim>
1257
+ <dim>3</dim>
1258
+ </port>
1259
+ </input>
1260
+ <output>
1261
+ <port id="1" precision="FP32">
1262
+ <dim>512</dim>
1263
+ <dim>512</dim>
1264
+ <dim>3</dim>
1265
+ <dim>3</dim>
1266
+ </port>
1267
+ </output>
1268
+ </layer>
1269
+ <layer id="65" name="Multiply_2546" type="Convolution" version="opset1">
1270
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
1271
+ <input>
1272
+ <port id="0" precision="FP32">
1273
+ <dim>1</dim>
1274
+ <dim>512</dim>
1275
+ <dim>12</dim>
1276
+ <dim>250</dim>
1277
+ </port>
1278
+ <port id="1" precision="FP32">
1279
+ <dim>512</dim>
1280
+ <dim>512</dim>
1281
+ <dim>3</dim>
1282
+ <dim>3</dim>
1283
+ </port>
1284
+ </input>
1285
+ <output>
1286
+ <port id="2" precision="FP32">
1287
+ <dim>1</dim>
1288
+ <dim>512</dim>
1289
+ <dim>12</dim>
1290
+ <dim>250</dim>
1291
+ </port>
1292
+ </output>
1293
+ </layer>
1294
+ <layer id="66" name="Constant_2551_compressed" type="Const" version="opset1">
1295
+ <data element_type="f16" shape="1, 512, 1, 1" offset="10547586" size="1024" />
1296
+ <output>
1297
+ <port id="0" precision="FP16">
1298
+ <dim>1</dim>
1299
+ <dim>512</dim>
1300
+ <dim>1</dim>
1301
+ <dim>1</dim>
1302
+ </port>
1303
+ </output>
1304
+ </layer>
1305
+ <layer id="67" name="Constant_2551" type="Convert" version="opset1">
1306
+ <data destination_type="f32" />
1307
+ <rt_info>
1308
+ <attribute name="decompression" version="0" />
1309
+ </rt_info>
1310
+ <input>
1311
+ <port id="0" precision="FP16">
1312
+ <dim>1</dim>
1313
+ <dim>512</dim>
1314
+ <dim>1</dim>
1315
+ <dim>1</dim>
1316
+ </port>
1317
+ </input>
1318
+ <output>
1319
+ <port id="1" precision="FP32">
1320
+ <dim>1</dim>
1321
+ <dim>512</dim>
1322
+ <dim>1</dim>
1323
+ <dim>1</dim>
1324
+ </port>
1325
+ </output>
1326
+ </layer>
1327
+ <layer id="68" name="124" type="Add" version="opset1">
1328
+ <data auto_broadcast="numpy" />
1329
+ <input>
1330
+ <port id="0" precision="FP32">
1331
+ <dim>1</dim>
1332
+ <dim>512</dim>
1333
+ <dim>12</dim>
1334
+ <dim>250</dim>
1335
+ </port>
1336
+ <port id="1" precision="FP32">
1337
+ <dim>1</dim>
1338
+ <dim>512</dim>
1339
+ <dim>1</dim>
1340
+ <dim>1</dim>
1341
+ </port>
1342
+ </input>
1343
+ <output>
1344
+ <port id="2" precision="FP32" names="124">
1345
+ <dim>1</dim>
1346
+ <dim>512</dim>
1347
+ <dim>12</dim>
1348
+ <dim>250</dim>
1349
+ </port>
1350
+ </output>
1351
+ </layer>
1352
+ <layer id="69" name="125" type="ReLU" version="opset1">
1353
+ <input>
1354
+ <port id="0" precision="FP32">
1355
+ <dim>1</dim>
1356
+ <dim>512</dim>
1357
+ <dim>12</dim>
1358
+ <dim>250</dim>
1359
+ </port>
1360
+ </input>
1361
+ <output>
1362
+ <port id="1" precision="FP32" names="125">
1363
+ <dim>1</dim>
1364
+ <dim>512</dim>
1365
+ <dim>12</dim>
1366
+ <dim>250</dim>
1367
+ </port>
1368
+ </output>
1369
+ </layer>
1370
+ <layer id="70" name="Multiply_2654_compressed" type="Const" version="opset1">
1371
+ <data element_type="f16" shape="512, 512, 3, 3" offset="10548610" size="4718592" />
1372
+ <output>
1373
+ <port id="0" precision="FP16">
1374
+ <dim>512</dim>
1375
+ <dim>512</dim>
1376
+ <dim>3</dim>
1377
+ <dim>3</dim>
1378
+ </port>
1379
+ </output>
1380
+ </layer>
1381
+ <layer id="71" name="Multiply_2654" type="Convert" version="opset1">
1382
+ <data destination_type="f32" />
1383
+ <rt_info>
1384
+ <attribute name="decompression" version="0" />
1385
+ </rt_info>
1386
+ <input>
1387
+ <port id="0" precision="FP16">
1388
+ <dim>512</dim>
1389
+ <dim>512</dim>
1390
+ <dim>3</dim>
1391
+ <dim>3</dim>
1392
+ </port>
1393
+ </input>
1394
+ <output>
1395
+ <port id="1" precision="FP32">
1396
+ <dim>512</dim>
1397
+ <dim>512</dim>
1398
+ <dim>3</dim>
1399
+ <dim>3</dim>
1400
+ </port>
1401
+ </output>
1402
+ </layer>
1403
+ <layer id="72" name="Multiply_2556" type="Convolution" version="opset1">
1404
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
1405
+ <input>
1406
+ <port id="0" precision="FP32">
1407
+ <dim>1</dim>
1408
+ <dim>512</dim>
1409
+ <dim>12</dim>
1410
+ <dim>250</dim>
1411
+ </port>
1412
+ <port id="1" precision="FP32">
1413
+ <dim>512</dim>
1414
+ <dim>512</dim>
1415
+ <dim>3</dim>
1416
+ <dim>3</dim>
1417
+ </port>
1418
+ </input>
1419
+ <output>
1420
+ <port id="2" precision="FP32">
1421
+ <dim>1</dim>
1422
+ <dim>512</dim>
1423
+ <dim>12</dim>
1424
+ <dim>250</dim>
1425
+ </port>
1426
+ </output>
1427
+ </layer>
1428
+ <layer id="73" name="Constant_2561_compressed" type="Const" version="opset1">
1429
+ <data element_type="f16" shape="1, 512, 1, 1" offset="15267202" size="1024" />
1430
+ <output>
1431
+ <port id="0" precision="FP16">
1432
+ <dim>1</dim>
1433
+ <dim>512</dim>
1434
+ <dim>1</dim>
1435
+ <dim>1</dim>
1436
+ </port>
1437
+ </output>
1438
+ </layer>
1439
+ <layer id="74" name="Constant_2561" type="Convert" version="opset1">
1440
+ <data destination_type="f32" />
1441
+ <rt_info>
1442
+ <attribute name="decompression" version="0" />
1443
+ </rt_info>
1444
+ <input>
1445
+ <port id="0" precision="FP16">
1446
+ <dim>1</dim>
1447
+ <dim>512</dim>
1448
+ <dim>1</dim>
1449
+ <dim>1</dim>
1450
+ </port>
1451
+ </input>
1452
+ <output>
1453
+ <port id="1" precision="FP32">
1454
+ <dim>1</dim>
1455
+ <dim>512</dim>
1456
+ <dim>1</dim>
1457
+ <dim>1</dim>
1458
+ </port>
1459
+ </output>
1460
+ </layer>
1461
+ <layer id="75" name="127" type="Add" version="opset1">
1462
+ <data auto_broadcast="numpy" />
1463
+ <input>
1464
+ <port id="0" precision="FP32">
1465
+ <dim>1</dim>
1466
+ <dim>512</dim>
1467
+ <dim>12</dim>
1468
+ <dim>250</dim>
1469
+ </port>
1470
+ <port id="1" precision="FP32">
1471
+ <dim>1</dim>
1472
+ <dim>512</dim>
1473
+ <dim>1</dim>
1474
+ <dim>1</dim>
1475
+ </port>
1476
+ </input>
1477
+ <output>
1478
+ <port id="2" precision="FP32" names="127">
1479
+ <dim>1</dim>
1480
+ <dim>512</dim>
1481
+ <dim>12</dim>
1482
+ <dim>250</dim>
1483
+ </port>
1484
+ </output>
1485
+ </layer>
1486
+ <layer id="76" name="128" type="ReLU" version="opset1">
1487
+ <input>
1488
+ <port id="0" precision="FP32">
1489
+ <dim>1</dim>
1490
+ <dim>512</dim>
1491
+ <dim>12</dim>
1492
+ <dim>250</dim>
1493
+ </port>
1494
+ </input>
1495
+ <output>
1496
+ <port id="1" precision="FP32" names="128">
1497
+ <dim>1</dim>
1498
+ <dim>512</dim>
1499
+ <dim>12</dim>
1500
+ <dim>250</dim>
1501
+ </port>
1502
+ </output>
1503
+ </layer>
1504
+ <layer id="77" name="130" type="MaxPool" version="opset8">
1505
+ <data strides="2, 2" dilations="1, 1" pads_begin="0, 0" pads_end="0, 0" kernel="2, 2" rounding_type="floor" auto_pad="explicit" index_element_type="i64" axis="0" />
1506
+ <input>
1507
+ <port id="0" precision="FP32">
1508
+ <dim>1</dim>
1509
+ <dim>512</dim>
1510
+ <dim>12</dim>
1511
+ <dim>250</dim>
1512
+ </port>
1513
+ </input>
1514
+ <output>
1515
+ <port id="1" precision="FP32" names="130">
1516
+ <dim>1</dim>
1517
+ <dim>512</dim>
1518
+ <dim>6</dim>
1519
+ <dim>125</dim>
1520
+ </port>
1521
+ <port id="2" precision="I64">
1522
+ <dim>1</dim>
1523
+ <dim>512</dim>
1524
+ <dim>6</dim>
1525
+ <dim>125</dim>
1526
+ </port>
1527
+ </output>
1528
+ </layer>
1529
+ <layer id="78" name="Multiply_2660_compressed" type="Const" version="opset1">
1530
+ <data element_type="f16" shape="512, 512, 3, 3" offset="15268226" size="4718592" />
1531
+ <output>
1532
+ <port id="0" precision="FP16">
1533
+ <dim>512</dim>
1534
+ <dim>512</dim>
1535
+ <dim>3</dim>
1536
+ <dim>3</dim>
1537
+ </port>
1538
+ </output>
1539
+ </layer>
1540
+ <layer id="79" name="Multiply_2660" type="Convert" version="opset1">
1541
+ <data destination_type="f32" />
1542
+ <rt_info>
1543
+ <attribute name="decompression" version="0" />
1544
+ </rt_info>
1545
+ <input>
1546
+ <port id="0" precision="FP16">
1547
+ <dim>512</dim>
1548
+ <dim>512</dim>
1549
+ <dim>3</dim>
1550
+ <dim>3</dim>
1551
+ </port>
1552
+ </input>
1553
+ <output>
1554
+ <port id="1" precision="FP32">
1555
+ <dim>512</dim>
1556
+ <dim>512</dim>
1557
+ <dim>3</dim>
1558
+ <dim>3</dim>
1559
+ </port>
1560
+ </output>
1561
+ </layer>
1562
+ <layer id="80" name="Multiply_2566" type="Convolution" version="opset1">
1563
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
1564
+ <input>
1565
+ <port id="0" precision="FP32">
1566
+ <dim>1</dim>
1567
+ <dim>512</dim>
1568
+ <dim>6</dim>
1569
+ <dim>125</dim>
1570
+ </port>
1571
+ <port id="1" precision="FP32">
1572
+ <dim>512</dim>
1573
+ <dim>512</dim>
1574
+ <dim>3</dim>
1575
+ <dim>3</dim>
1576
+ </port>
1577
+ </input>
1578
+ <output>
1579
+ <port id="2" precision="FP32">
1580
+ <dim>1</dim>
1581
+ <dim>512</dim>
1582
+ <dim>6</dim>
1583
+ <dim>125</dim>
1584
+ </port>
1585
+ </output>
1586
+ </layer>
1587
+ <layer id="81" name="Constant_2571_compressed" type="Const" version="opset1">
1588
+ <data element_type="f16" shape="1, 512, 1, 1" offset="19986818" size="1024" />
1589
+ <output>
1590
+ <port id="0" precision="FP16">
1591
+ <dim>1</dim>
1592
+ <dim>512</dim>
1593
+ <dim>1</dim>
1594
+ <dim>1</dim>
1595
+ </port>
1596
+ </output>
1597
+ </layer>
1598
+ <layer id="82" name="Constant_2571" type="Convert" version="opset1">
1599
+ <data destination_type="f32" />
1600
+ <rt_info>
1601
+ <attribute name="decompression" version="0" />
1602
+ </rt_info>
1603
+ <input>
1604
+ <port id="0" precision="FP16">
1605
+ <dim>1</dim>
1606
+ <dim>512</dim>
1607
+ <dim>1</dim>
1608
+ <dim>1</dim>
1609
+ </port>
1610
+ </input>
1611
+ <output>
1612
+ <port id="1" precision="FP32">
1613
+ <dim>1</dim>
1614
+ <dim>512</dim>
1615
+ <dim>1</dim>
1616
+ <dim>1</dim>
1617
+ </port>
1618
+ </output>
1619
+ </layer>
1620
+ <layer id="83" name="133" type="Add" version="opset1">
1621
+ <data auto_broadcast="numpy" />
1622
+ <input>
1623
+ <port id="0" precision="FP32">
1624
+ <dim>1</dim>
1625
+ <dim>512</dim>
1626
+ <dim>6</dim>
1627
+ <dim>125</dim>
1628
+ </port>
1629
+ <port id="1" precision="FP32">
1630
+ <dim>1</dim>
1631
+ <dim>512</dim>
1632
+ <dim>1</dim>
1633
+ <dim>1</dim>
1634
+ </port>
1635
+ </input>
1636
+ <output>
1637
+ <port id="2" precision="FP32" names="133">
1638
+ <dim>1</dim>
1639
+ <dim>512</dim>
1640
+ <dim>6</dim>
1641
+ <dim>125</dim>
1642
+ </port>
1643
+ </output>
1644
+ </layer>
1645
+ <layer id="84" name="134" type="ReLU" version="opset1">
1646
+ <input>
1647
+ <port id="0" precision="FP32">
1648
+ <dim>1</dim>
1649
+ <dim>512</dim>
1650
+ <dim>6</dim>
1651
+ <dim>125</dim>
1652
+ </port>
1653
+ </input>
1654
+ <output>
1655
+ <port id="1" precision="FP32" names="134">
1656
+ <dim>1</dim>
1657
+ <dim>512</dim>
1658
+ <dim>6</dim>
1659
+ <dim>125</dim>
1660
+ </port>
1661
+ </output>
1662
+ </layer>
1663
+ <layer id="85" name="Multiply_2666_compressed" type="Const" version="opset1">
1664
+ <data element_type="f16" shape="512, 512, 3, 3" offset="19987842" size="4718592" />
1665
+ <output>
1666
+ <port id="0" precision="FP16">
1667
+ <dim>512</dim>
1668
+ <dim>512</dim>
1669
+ <dim>3</dim>
1670
+ <dim>3</dim>
1671
+ </port>
1672
+ </output>
1673
+ </layer>
1674
+ <layer id="86" name="Multiply_2666" type="Convert" version="opset1">
1675
+ <data destination_type="f32" />
1676
+ <rt_info>
1677
+ <attribute name="decompression" version="0" />
1678
+ </rt_info>
1679
+ <input>
1680
+ <port id="0" precision="FP16">
1681
+ <dim>512</dim>
1682
+ <dim>512</dim>
1683
+ <dim>3</dim>
1684
+ <dim>3</dim>
1685
+ </port>
1686
+ </input>
1687
+ <output>
1688
+ <port id="1" precision="FP32">
1689
+ <dim>512</dim>
1690
+ <dim>512</dim>
1691
+ <dim>3</dim>
1692
+ <dim>3</dim>
1693
+ </port>
1694
+ </output>
1695
+ </layer>
1696
+ <layer id="87" name="Multiply_2576" type="Convolution" version="opset1">
1697
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
1698
+ <input>
1699
+ <port id="0" precision="FP32">
1700
+ <dim>1</dim>
1701
+ <dim>512</dim>
1702
+ <dim>6</dim>
1703
+ <dim>125</dim>
1704
+ </port>
1705
+ <port id="1" precision="FP32">
1706
+ <dim>512</dim>
1707
+ <dim>512</dim>
1708
+ <dim>3</dim>
1709
+ <dim>3</dim>
1710
+ </port>
1711
+ </input>
1712
+ <output>
1713
+ <port id="2" precision="FP32">
1714
+ <dim>1</dim>
1715
+ <dim>512</dim>
1716
+ <dim>6</dim>
1717
+ <dim>125</dim>
1718
+ </port>
1719
+ </output>
1720
+ </layer>
1721
+ <layer id="88" name="Constant_2581_compressed" type="Const" version="opset1">
1722
+ <data element_type="f16" shape="1, 512, 1, 1" offset="24706434" size="1024" />
1723
+ <output>
1724
+ <port id="0" precision="FP16">
1725
+ <dim>1</dim>
1726
+ <dim>512</dim>
1727
+ <dim>1</dim>
1728
+ <dim>1</dim>
1729
+ </port>
1730
+ </output>
1731
+ </layer>
1732
+ <layer id="89" name="Constant_2581" type="Convert" version="opset1">
1733
+ <data destination_type="f32" />
1734
+ <rt_info>
1735
+ <attribute name="decompression" version="0" />
1736
+ </rt_info>
1737
+ <input>
1738
+ <port id="0" precision="FP16">
1739
+ <dim>1</dim>
1740
+ <dim>512</dim>
1741
+ <dim>1</dim>
1742
+ <dim>1</dim>
1743
+ </port>
1744
+ </input>
1745
+ <output>
1746
+ <port id="1" precision="FP32">
1747
+ <dim>1</dim>
1748
+ <dim>512</dim>
1749
+ <dim>1</dim>
1750
+ <dim>1</dim>
1751
+ </port>
1752
+ </output>
1753
+ </layer>
1754
+ <layer id="90" name="136" type="Add" version="opset1">
1755
+ <data auto_broadcast="numpy" />
1756
+ <input>
1757
+ <port id="0" precision="FP32">
1758
+ <dim>1</dim>
1759
+ <dim>512</dim>
1760
+ <dim>6</dim>
1761
+ <dim>125</dim>
1762
+ </port>
1763
+ <port id="1" precision="FP32">
1764
+ <dim>1</dim>
1765
+ <dim>512</dim>
1766
+ <dim>1</dim>
1767
+ <dim>1</dim>
1768
+ </port>
1769
+ </input>
1770
+ <output>
1771
+ <port id="2" precision="FP32" names="136">
1772
+ <dim>1</dim>
1773
+ <dim>512</dim>
1774
+ <dim>6</dim>
1775
+ <dim>125</dim>
1776
+ </port>
1777
+ </output>
1778
+ </layer>
1779
+ <layer id="91" name="137" type="ReLU" version="opset1">
1780
+ <input>
1781
+ <port id="0" precision="FP32">
1782
+ <dim>1</dim>
1783
+ <dim>512</dim>
1784
+ <dim>6</dim>
1785
+ <dim>125</dim>
1786
+ </port>
1787
+ </input>
1788
+ <output>
1789
+ <port id="1" precision="FP32" names="137">
1790
+ <dim>1</dim>
1791
+ <dim>512</dim>
1792
+ <dim>6</dim>
1793
+ <dim>125</dim>
1794
+ </port>
1795
+ </output>
1796
+ </layer>
1797
+ <layer id="92" name="Multiply_2672_compressed" type="Const" version="opset1">
1798
+ <data element_type="f16" shape="512, 512, 3, 3" offset="24707458" size="4718592" />
1799
+ <output>
1800
+ <port id="0" precision="FP16">
1801
+ <dim>512</dim>
1802
+ <dim>512</dim>
1803
+ <dim>3</dim>
1804
+ <dim>3</dim>
1805
+ </port>
1806
+ </output>
1807
+ </layer>
1808
+ <layer id="93" name="Multiply_2672" type="Convert" version="opset1">
1809
+ <data destination_type="f32" />
1810
+ <rt_info>
1811
+ <attribute name="decompression" version="0" />
1812
+ </rt_info>
1813
+ <input>
1814
+ <port id="0" precision="FP16">
1815
+ <dim>512</dim>
1816
+ <dim>512</dim>
1817
+ <dim>3</dim>
1818
+ <dim>3</dim>
1819
+ </port>
1820
+ </input>
1821
+ <output>
1822
+ <port id="1" precision="FP32">
1823
+ <dim>512</dim>
1824
+ <dim>512</dim>
1825
+ <dim>3</dim>
1826
+ <dim>3</dim>
1827
+ </port>
1828
+ </output>
1829
+ </layer>
1830
+ <layer id="94" name="Multiply_2586" type="Convolution" version="opset1">
1831
+ <data strides="1, 1" dilations="1, 1" pads_begin="1, 1" pads_end="1, 1" auto_pad="explicit" />
1832
+ <input>
1833
+ <port id="0" precision="FP32">
1834
+ <dim>1</dim>
1835
+ <dim>512</dim>
1836
+ <dim>6</dim>
1837
+ <dim>125</dim>
1838
+ </port>
1839
+ <port id="1" precision="FP32">
1840
+ <dim>512</dim>
1841
+ <dim>512</dim>
1842
+ <dim>3</dim>
1843
+ <dim>3</dim>
1844
+ </port>
1845
+ </input>
1846
+ <output>
1847
+ <port id="2" precision="FP32">
1848
+ <dim>1</dim>
1849
+ <dim>512</dim>
1850
+ <dim>6</dim>
1851
+ <dim>125</dim>
1852
+ </port>
1853
+ </output>
1854
+ </layer>
1855
+ <layer id="95" name="Constant_2591_compressed" type="Const" version="opset1">
1856
+ <data element_type="f16" shape="1, 512, 1, 1" offset="29426050" size="1024" />
1857
+ <output>
1858
+ <port id="0" precision="FP16">
1859
+ <dim>1</dim>
1860
+ <dim>512</dim>
1861
+ <dim>1</dim>
1862
+ <dim>1</dim>
1863
+ </port>
1864
+ </output>
1865
+ </layer>
1866
+ <layer id="96" name="Constant_2591" type="Convert" version="opset1">
1867
+ <data destination_type="f32" />
1868
+ <rt_info>
1869
+ <attribute name="decompression" version="0" />
1870
+ </rt_info>
1871
+ <input>
1872
+ <port id="0" precision="FP16">
1873
+ <dim>1</dim>
1874
+ <dim>512</dim>
1875
+ <dim>1</dim>
1876
+ <dim>1</dim>
1877
+ </port>
1878
+ </input>
1879
+ <output>
1880
+ <port id="1" precision="FP32">
1881
+ <dim>1</dim>
1882
+ <dim>512</dim>
1883
+ <dim>1</dim>
1884
+ <dim>1</dim>
1885
+ </port>
1886
+ </output>
1887
+ </layer>
1888
+ <layer id="97" name="139" type="Add" version="opset1">
1889
+ <data auto_broadcast="numpy" />
1890
+ <input>
1891
+ <port id="0" precision="FP32">
1892
+ <dim>1</dim>
1893
+ <dim>512</dim>
1894
+ <dim>6</dim>
1895
+ <dim>125</dim>
1896
+ </port>
1897
+ <port id="1" precision="FP32">
1898
+ <dim>1</dim>
1899
+ <dim>512</dim>
1900
+ <dim>1</dim>
1901
+ <dim>1</dim>
1902
+ </port>
1903
+ </input>
1904
+ <output>
1905
+ <port id="2" precision="FP32" names="139">
1906
+ <dim>1</dim>
1907
+ <dim>512</dim>
1908
+ <dim>6</dim>
1909
+ <dim>125</dim>
1910
+ </port>
1911
+ </output>
1912
+ </layer>
1913
+ <layer id="98" name="140" type="ReLU" version="opset1">
1914
+ <input>
1915
+ <port id="0" precision="FP32">
1916
+ <dim>1</dim>
1917
+ <dim>512</dim>
1918
+ <dim>6</dim>
1919
+ <dim>125</dim>
1920
+ </port>
1921
+ </input>
1922
+ <output>
1923
+ <port id="1" precision="FP32" names="140">
1924
+ <dim>1</dim>
1925
+ <dim>512</dim>
1926
+ <dim>6</dim>
1927
+ <dim>125</dim>
1928
+ </port>
1929
+ </output>
1930
+ </layer>
1931
+ <layer id="99" name="142" type="MaxPool" version="opset8">
1932
+ <data strides="2, 2" dilations="1, 1" pads_begin="0, 0" pads_end="0, 0" kernel="2, 2" rounding_type="floor" auto_pad="explicit" index_element_type="i64" axis="0" />
1933
+ <input>
1934
+ <port id="0" precision="FP32">
1935
+ <dim>1</dim>
1936
+ <dim>512</dim>
1937
+ <dim>6</dim>
1938
+ <dim>125</dim>
1939
+ </port>
1940
+ </input>
1941
+ <output>
1942
+ <port id="1" precision="FP32" names="142">
1943
+ <dim>1</dim>
1944
+ <dim>512</dim>
1945
+ <dim>3</dim>
1946
+ <dim>62</dim>
1947
+ </port>
1948
+ <port id="2" precision="I64">
1949
+ <dim>1</dim>
1950
+ <dim>512</dim>
1951
+ <dim>3</dim>
1952
+ <dim>62</dim>
1953
+ </port>
1954
+ </output>
1955
+ </layer>
1956
+ <layer id="100" name="144" type="Const" version="opset1">
1957
+ <data element_type="i64" shape="3" offset="29427074" size="24" />
1958
+ <output>
1959
+ <port id="0" precision="I64" names="144">
1960
+ <dim>3</dim>
1961
+ </port>
1962
+ </output>
1963
+ </layer>
1964
+ <layer id="101" name="145" type="Reshape" version="opset1">
1965
+ <data special_zero="true" />
1966
+ <input>
1967
+ <port id="0" precision="FP32">
1968
+ <dim>1</dim>
1969
+ <dim>512</dim>
1970
+ <dim>3</dim>
1971
+ <dim>62</dim>
1972
+ </port>
1973
+ <port id="1" precision="I64">
1974
+ <dim>3</dim>
1975
+ </port>
1976
+ </input>
1977
+ <output>
1978
+ <port id="2" precision="FP32" names="145">
1979
+ <dim>1</dim>
1980
+ <dim>512</dim>
1981
+ <dim>186</dim>
1982
+ </port>
1983
+ </output>
1984
+ </layer>
1985
+ <layer id="102" name="92_compressed" type="Const" version="opset1">
1986
+ <data element_type="f16" shape="4442, 512" offset="29427098" size="4548608" />
1987
+ <output>
1988
+ <port id="0" precision="FP16">
1989
+ <dim>4442</dim>
1990
+ <dim>512</dim>
1991
+ </port>
1992
+ </output>
1993
+ </layer>
1994
+ <layer id="103" name="92" type="Convert" version="opset1">
1995
+ <data destination_type="f32" />
1996
+ <rt_info>
1997
+ <attribute name="decompression" version="0" />
1998
+ </rt_info>
1999
+ <input>
2000
+ <port id="0" precision="FP16">
2001
+ <dim>4442</dim>
2002
+ <dim>512</dim>
2003
+ </port>
2004
+ </input>
2005
+ <output>
2006
+ <port id="1" precision="FP32" names="92">
2007
+ <dim>4442</dim>
2008
+ <dim>512</dim>
2009
+ </port>
2010
+ </output>
2011
+ </layer>
2012
+ <layer id="104" name="148" type="MatMul" version="opset1">
2013
+ <data transpose_a="true" transpose_b="true" />
2014
+ <input>
2015
+ <port id="0" precision="FP32">
2016
+ <dim>1</dim>
2017
+ <dim>512</dim>
2018
+ <dim>186</dim>
2019
+ </port>
2020
+ <port id="1" precision="FP32">
2021
+ <dim>4442</dim>
2022
+ <dim>512</dim>
2023
+ </port>
2024
+ </input>
2025
+ <output>
2026
+ <port id="2" precision="FP32" names="148">
2027
+ <dim>1</dim>
2028
+ <dim>186</dim>
2029
+ <dim>4442</dim>
2030
+ </port>
2031
+ </output>
2032
+ </layer>
2033
+ <layer id="105" name="Constant_2752_compressed" type="Const" version="opset1">
2034
+ <data element_type="f16" shape="1, 1, 4442" offset="33975706" size="8884" />
2035
+ <output>
2036
+ <port id="0" precision="FP16">
2037
+ <dim>1</dim>
2038
+ <dim>1</dim>
2039
+ <dim>4442</dim>
2040
+ </port>
2041
+ </output>
2042
+ </layer>
2043
+ <layer id="106" name="Constant_2752" type="Convert" version="opset1">
2044
+ <data destination_type="f32" />
2045
+ <rt_info>
2046
+ <attribute name="decompression" version="0" />
2047
+ </rt_info>
2048
+ <input>
2049
+ <port id="0" precision="FP16">
2050
+ <dim>1</dim>
2051
+ <dim>1</dim>
2052
+ <dim>4442</dim>
2053
+ </port>
2054
+ </input>
2055
+ <output>
2056
+ <port id="1" precision="FP32">
2057
+ <dim>1</dim>
2058
+ <dim>1</dim>
2059
+ <dim>4442</dim>
2060
+ </port>
2061
+ </output>
2062
+ </layer>
2063
+ <layer id="107" name="149" type="Add" version="opset1">
2064
+ <data auto_broadcast="numpy" />
2065
+ <input>
2066
+ <port id="0" precision="FP32">
2067
+ <dim>1</dim>
2068
+ <dim>186</dim>
2069
+ <dim>4442</dim>
2070
+ </port>
2071
+ <port id="1" precision="FP32">
2072
+ <dim>1</dim>
2073
+ <dim>1</dim>
2074
+ <dim>4442</dim>
2075
+ </port>
2076
+ </input>
2077
+ <output>
2078
+ <port id="2" precision="FP32" names="149">
2079
+ <dim>1</dim>
2080
+ <dim>186</dim>
2081
+ <dim>4442</dim>
2082
+ </port>
2083
+ </output>
2084
+ </layer>
2085
+ <layer id="108" name="Constant_1597" type="Const" version="opset1">
2086
+ <data element_type="i64" shape="3" offset="33984590" size="24" />
2087
+ <output>
2088
+ <port id="0" precision="I64">
2089
+ <dim>3</dim>
2090
+ </port>
2091
+ </output>
2092
+ </layer>
2093
+ <layer id="109" name="output" type="Reshape" version="opset1">
2094
+ <data special_zero="true" />
2095
+ <input>
2096
+ <port id="0" precision="FP32">
2097
+ <dim>1</dim>
2098
+ <dim>186</dim>
2099
+ <dim>4442</dim>
2100
+ </port>
2101
+ <port id="1" precision="I64">
2102
+ <dim>3</dim>
2103
+ </port>
2104
+ </input>
2105
+ <output>
2106
+ <port id="2" precision="FP32" names="output">
2107
+ <dim>186</dim>
2108
+ <dim>1</dim>
2109
+ <dim>4442</dim>
2110
+ </port>
2111
+ </output>
2112
+ </layer>
2113
+ <layer id="110" name="output/sink_port_0" type="Result" version="opset1">
2114
+ <input>
2115
+ <port id="0" precision="FP32">
2116
+ <dim>186</dim>
2117
+ <dim>1</dim>
2118
+ <dim>4442</dim>
2119
+ </port>
2120
+ </input>
2121
+ </layer>
2122
+ </layers>
2123
+ <edges>
2124
+ <edge from-layer="0" from-port="0" to-layer="3" to-port="0" />
2125
+ <edge from-layer="1" from-port="0" to-layer="2" to-port="0" />
2126
+ <edge from-layer="2" from-port="1" to-layer="3" to-port="1" />
2127
+ <edge from-layer="3" from-port="2" to-layer="6" to-port="0" />
2128
+ <edge from-layer="4" from-port="0" to-layer="5" to-port="0" />
2129
+ <edge from-layer="5" from-port="1" to-layer="6" to-port="1" />
2130
+ <edge from-layer="6" from-port="2" to-layer="9" to-port="0" />
2131
+ <edge from-layer="7" from-port="0" to-layer="8" to-port="0" />
2132
+ <edge from-layer="8" from-port="1" to-layer="9" to-port="1" />
2133
+ <edge from-layer="9" from-port="2" to-layer="10" to-port="0" />
2134
+ <edge from-layer="10" from-port="1" to-layer="13" to-port="0" />
2135
+ <edge from-layer="11" from-port="0" to-layer="12" to-port="0" />
2136
+ <edge from-layer="12" from-port="1" to-layer="13" to-port="1" />
2137
+ <edge from-layer="13" from-port="2" to-layer="16" to-port="0" />
2138
+ <edge from-layer="14" from-port="0" to-layer="15" to-port="0" />
2139
+ <edge from-layer="15" from-port="1" to-layer="16" to-port="1" />
2140
+ <edge from-layer="16" from-port="2" to-layer="17" to-port="0" />
2141
+ <edge from-layer="17" from-port="1" to-layer="18" to-port="0" />
2142
+ <edge from-layer="18" from-port="1" to-layer="21" to-port="0" />
2143
+ <edge from-layer="19" from-port="0" to-layer="20" to-port="0" />
2144
+ <edge from-layer="20" from-port="1" to-layer="21" to-port="1" />
2145
+ <edge from-layer="21" from-port="2" to-layer="24" to-port="0" />
2146
+ <edge from-layer="22" from-port="0" to-layer="23" to-port="0" />
2147
+ <edge from-layer="23" from-port="1" to-layer="24" to-port="1" />
2148
+ <edge from-layer="24" from-port="2" to-layer="25" to-port="0" />
2149
+ <edge from-layer="25" from-port="1" to-layer="28" to-port="0" />
2150
+ <edge from-layer="26" from-port="0" to-layer="27" to-port="0" />
2151
+ <edge from-layer="27" from-port="1" to-layer="28" to-port="1" />
2152
+ <edge from-layer="28" from-port="2" to-layer="31" to-port="0" />
2153
+ <edge from-layer="29" from-port="0" to-layer="30" to-port="0" />
2154
+ <edge from-layer="30" from-port="1" to-layer="31" to-port="1" />
2155
+ <edge from-layer="31" from-port="2" to-layer="32" to-port="0" />
2156
+ <edge from-layer="32" from-port="1" to-layer="33" to-port="0" />
2157
+ <edge from-layer="33" from-port="1" to-layer="36" to-port="0" />
2158
+ <edge from-layer="34" from-port="0" to-layer="35" to-port="0" />
2159
+ <edge from-layer="35" from-port="1" to-layer="36" to-port="1" />
2160
+ <edge from-layer="36" from-port="2" to-layer="39" to-port="0" />
2161
+ <edge from-layer="37" from-port="0" to-layer="38" to-port="0" />
2162
+ <edge from-layer="38" from-port="1" to-layer="39" to-port="1" />
2163
+ <edge from-layer="39" from-port="2" to-layer="40" to-port="0" />
2164
+ <edge from-layer="40" from-port="1" to-layer="43" to-port="0" />
2165
+ <edge from-layer="41" from-port="0" to-layer="42" to-port="0" />
2166
+ <edge from-layer="42" from-port="1" to-layer="43" to-port="1" />
2167
+ <edge from-layer="43" from-port="2" to-layer="46" to-port="0" />
2168
+ <edge from-layer="44" from-port="0" to-layer="45" to-port="0" />
2169
+ <edge from-layer="45" from-port="1" to-layer="46" to-port="1" />
2170
+ <edge from-layer="46" from-port="2" to-layer="47" to-port="0" />
2171
+ <edge from-layer="47" from-port="1" to-layer="50" to-port="0" />
2172
+ <edge from-layer="48" from-port="0" to-layer="49" to-port="0" />
2173
+ <edge from-layer="49" from-port="1" to-layer="50" to-port="1" />
2174
+ <edge from-layer="50" from-port="2" to-layer="53" to-port="0" />
2175
+ <edge from-layer="51" from-port="0" to-layer="52" to-port="0" />
2176
+ <edge from-layer="52" from-port="1" to-layer="53" to-port="1" />
2177
+ <edge from-layer="53" from-port="2" to-layer="54" to-port="0" />
2178
+ <edge from-layer="54" from-port="1" to-layer="55" to-port="0" />
2179
+ <edge from-layer="55" from-port="1" to-layer="58" to-port="0" />
2180
+ <edge from-layer="56" from-port="0" to-layer="57" to-port="0" />
2181
+ <edge from-layer="57" from-port="1" to-layer="58" to-port="1" />
2182
+ <edge from-layer="58" from-port="2" to-layer="61" to-port="0" />
2183
+ <edge from-layer="59" from-port="0" to-layer="60" to-port="0" />
2184
+ <edge from-layer="60" from-port="1" to-layer="61" to-port="1" />
2185
+ <edge from-layer="61" from-port="2" to-layer="62" to-port="0" />
2186
+ <edge from-layer="62" from-port="1" to-layer="65" to-port="0" />
2187
+ <edge from-layer="63" from-port="0" to-layer="64" to-port="0" />
2188
+ <edge from-layer="64" from-port="1" to-layer="65" to-port="1" />
2189
+ <edge from-layer="65" from-port="2" to-layer="68" to-port="0" />
2190
+ <edge from-layer="66" from-port="0" to-layer="67" to-port="0" />
2191
+ <edge from-layer="67" from-port="1" to-layer="68" to-port="1" />
2192
+ <edge from-layer="68" from-port="2" to-layer="69" to-port="0" />
2193
+ <edge from-layer="69" from-port="1" to-layer="72" to-port="0" />
2194
+ <edge from-layer="70" from-port="0" to-layer="71" to-port="0" />
2195
+ <edge from-layer="71" from-port="1" to-layer="72" to-port="1" />
2196
+ <edge from-layer="72" from-port="2" to-layer="75" to-port="0" />
2197
+ <edge from-layer="73" from-port="0" to-layer="74" to-port="0" />
2198
+ <edge from-layer="74" from-port="1" to-layer="75" to-port="1" />
2199
+ <edge from-layer="75" from-port="2" to-layer="76" to-port="0" />
2200
+ <edge from-layer="76" from-port="1" to-layer="77" to-port="0" />
2201
+ <edge from-layer="77" from-port="1" to-layer="80" to-port="0" />
2202
+ <edge from-layer="78" from-port="0" to-layer="79" to-port="0" />
2203
+ <edge from-layer="79" from-port="1" to-layer="80" to-port="1" />
2204
+ <edge from-layer="80" from-port="2" to-layer="83" to-port="0" />
2205
+ <edge from-layer="81" from-port="0" to-layer="82" to-port="0" />
2206
+ <edge from-layer="82" from-port="1" to-layer="83" to-port="1" />
2207
+ <edge from-layer="83" from-port="2" to-layer="84" to-port="0" />
2208
+ <edge from-layer="84" from-port="1" to-layer="87" to-port="0" />
2209
+ <edge from-layer="85" from-port="0" to-layer="86" to-port="0" />
2210
+ <edge from-layer="86" from-port="1" to-layer="87" to-port="1" />
2211
+ <edge from-layer="87" from-port="2" to-layer="90" to-port="0" />
2212
+ <edge from-layer="88" from-port="0" to-layer="89" to-port="0" />
2213
+ <edge from-layer="89" from-port="1" to-layer="90" to-port="1" />
2214
+ <edge from-layer="90" from-port="2" to-layer="91" to-port="0" />
2215
+ <edge from-layer="91" from-port="1" to-layer="94" to-port="0" />
2216
+ <edge from-layer="92" from-port="0" to-layer="93" to-port="0" />
2217
+ <edge from-layer="93" from-port="1" to-layer="94" to-port="1" />
2218
+ <edge from-layer="94" from-port="2" to-layer="97" to-port="0" />
2219
+ <edge from-layer="95" from-port="0" to-layer="96" to-port="0" />
2220
+ <edge from-layer="96" from-port="1" to-layer="97" to-port="1" />
2221
+ <edge from-layer="97" from-port="2" to-layer="98" to-port="0" />
2222
+ <edge from-layer="98" from-port="1" to-layer="99" to-port="0" />
2223
+ <edge from-layer="99" from-port="1" to-layer="101" to-port="0" />
2224
+ <edge from-layer="100" from-port="0" to-layer="101" to-port="1" />
2225
+ <edge from-layer="101" from-port="2" to-layer="104" to-port="0" />
2226
+ <edge from-layer="102" from-port="0" to-layer="103" to-port="0" />
2227
+ <edge from-layer="103" from-port="1" to-layer="104" to-port="1" />
2228
+ <edge from-layer="104" from-port="2" to-layer="107" to-port="0" />
2229
+ <edge from-layer="105" from-port="0" to-layer="106" to-port="0" />
2230
+ <edge from-layer="106" from-port="1" to-layer="107" to-port="1" />
2231
+ <edge from-layer="107" from-port="2" to-layer="109" to-port="0" />
2232
+ <edge from-layer="108" from-port="0" to-layer="109" to-port="1" />
2233
+ <edge from-layer="109" from-port="2" to-layer="110" to-port="0" />
2234
+ </edges>
2235
+ <rt_info>
2236
+ <MO_version value="custom_HEAD_f6ee6e92f846a8c665e4a7089c51481f9689a3b5" />
2237
+ <Runtime_version value="2023.0.0-10521-f6ee6e92f84-HEAD" />
2238
+ <conversion_parameters>
2239
+ <framework value="onnx" />
2240
+ <input value="actual_input" />
2241
+ <input_model value="DIR/handwritten-japanese-recognition-0001.onnx" />
2242
+ <input_shape value="[1,1,96,2000]" />
2243
+ <layout value="actual_input(nchw)" />
2244
+ <mean_values value="actual_input[127.5]" />
2245
+ <model_name value="handwritten-japanese-recognition-0001" />
2246
+ <output value="output" />
2247
+ <output_dir value="DIR" />
2248
+ <scale_values value="actual_input[127.5]" />
2249
+ </conversion_parameters>
2250
+ <legacy_frontend value="False" />
2251
+ </rt_info>
2252
+ </net>
pyproject.toml ADDED
@@ -0,0 +1,21 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [tool.black]
2
+ line-length = 79
3
+ include = '\.pyi?$'
4
+ exclude = '''
5
+ /(
6
+ \.git
7
+ | \.idea
8
+ | \.pytest_cache
9
+ | \.tox
10
+ | \.venv
11
+ | _build
12
+ | buck-out
13
+ | build
14
+ | dist
15
+ )/
16
+ '''
17
+
18
+ [tool.isort]
19
+ profile = "black"
20
+ force_grid_wrap=2
21
+ multi_line_output=3
requirements.txt ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ streamlit
2
+ openvino
3
+ opencv-python
4
+ gradio
src/__Init__.py ADDED
File without changes
src/config.py ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #from streamlit_webrtc import RTCConfiguration
2
+ from collections import namedtuple
3
+
4
+ # MODELS
5
+ MODEL_DIR = "models/handwritten-japanese-recognition-0001.xml"
6
+ #MODEL_NAME = "handwritten-japanese-recognition-0001"
7
+ DATA_DIR = "data"
8
+ DICT_DIR = "dict/japanese_charlist.txt"
9
+ DEVICE = "CPU"
10
+
11
+ # Precision used by the model.
12
+ precision = "FP16"
13
+
14
+ # Config server for webcam
15
+ #RTC_CONFIGURATION = RTCConfiguration(
16
+ # {"iceServers": [{"urls": ["stun:stun4.l.google.com:19302"]}]},
17
+ #)
18
+
19
+ # IMAGE
20
+ IMAGE_EXAMPLE = "data/in_1.png"
21
+ IMAGE_TYPES = ['jpg','png','jpeg','JPG']
22
+
23
+
24
+ Language = namedtuple(
25
+ typename="Language", field_names=["model_name", "dict_name", "demo_image_name"]
26
+ )
src/detect_mask_image.py ADDED
@@ -0,0 +1,225 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # USAGE
2
+ # python detect_mask_image.py --image images/pic1.jpeg
3
+
4
+ import argparse
5
+ import os
6
+
7
+ import cv2
8
+ import numpy as np
9
+
10
+ # import the necessary packages
11
+ from tensorflow.keras.applications.mobilenet_v2 import preprocess_input
12
+ from tensorflow.keras.models import load_model
13
+ from tensorflow.keras.preprocessing.image import img_to_array
14
+
15
+
16
+ def mask_image():
17
+ # construct the argument parser and parse the arguments
18
+ ap = argparse.ArgumentParser()
19
+ ap.add_argument(
20
+ "-i",
21
+ "--image",
22
+ required=True,
23
+ help="path to input image",
24
+ )
25
+ ap.add_argument(
26
+ "-f",
27
+ "--face",
28
+ type=str,
29
+ default="face_detector",
30
+ help="path to face detector model directory",
31
+ )
32
+ ap.add_argument(
33
+ "-m",
34
+ "--model",
35
+ type=str,
36
+ default="mask_detector.model",
37
+ help="path to trained face mask detector model",
38
+ )
39
+ ap.add_argument(
40
+ "-c",
41
+ "--confidence",
42
+ type=float,
43
+ default=0.5,
44
+ help="minimum probability to filter weak detections",
45
+ )
46
+ args = vars(ap.parse_args())
47
+
48
+ # load our serialized face detector model from disk
49
+ print("[INFO] loading face detector model...")
50
+ prototxtPath = os.path.sep.join([args["face"], "deploy.prototxt"])
51
+ weightsPath = os.path.sep.join(
52
+ [
53
+ args["face"],
54
+ "res10_300x300_ssd_iter_140000.caffemodel",
55
+ ],
56
+ )
57
+ net = cv2.dnn.readNet(prototxtPath, weightsPath)
58
+
59
+ # load the face mask detector model from disk
60
+ print("[INFO] loading face mask detector model...")
61
+ model = load_model(args["model"])
62
+
63
+ # load the input image from disk, clone it, and grab the image spatial
64
+ # dimensions
65
+ image = cv2.imread(args["image"])
66
+ (h, w) = image.shape[:2]
67
+
68
+ # construct a blob from the image
69
+ blob = cv2.dnn.blobFromImage(
70
+ image,
71
+ 1.0,
72
+ (300, 300),
73
+ (104.0, 177.0, 123.0),
74
+ )
75
+
76
+ # pass the blob through the network and obtain the face detections
77
+ print("[INFO] computing face detections...")
78
+ net.setInput(blob)
79
+ detections = net.forward()
80
+
81
+ # loop over the detections
82
+ for i in range(0, detections.shape[2]):
83
+ # extract the confidence (i.e., probability) associated with
84
+ # the detection
85
+ confidence = detections[0, 0, i, 2]
86
+
87
+ # filter out weak detections by ensuring the confidence is
88
+ # greater than the minimum confidence
89
+ if confidence > args["confidence"]:
90
+ # compute the (x, y)-coordinates of the bounding box for
91
+ # the object
92
+ box = detections[0, 0, i, 3:7] * np.array([w, h, w, h])
93
+ (startX, startY, endX, endY) = box.astype("int")
94
+
95
+ # ensure the bounding boxes fall within the dimensions of
96
+ # the frame
97
+ (startX, startY) = (max(0, startX), max(0, startY))
98
+ (endX, endY) = (min(w - 1, endX), min(h - 1, endY))
99
+
100
+ # extract the face ROI, convert it from BGR to RGB channel
101
+ # ordering, resize it to 224x224, and preprocess it
102
+ face = image[startY:endY, startX:endX]
103
+ face = cv2.cvtColor(face, cv2.COLOR_BGR2RGB)
104
+ face = cv2.resize(face, (224, 224))
105
+ face = img_to_array(face)
106
+ face = preprocess_input(face)
107
+ face = np.expand_dims(face, axis=0)
108
+
109
+ # pass the face through the model to determine if the face
110
+ # has a mask or not
111
+ (mask, withoutMask) = model.predict(face)[0]
112
+
113
+ # determine the class label and color we'll use to draw
114
+ # the bounding box and text
115
+ label = "Mask" if mask > withoutMask else "No Mask"
116
+ color = (0, 255, 0) if label == "Mask" else (0, 0, 255)
117
+
118
+ # include the probability in the label
119
+ label = f"{label}: {max(mask, withoutMask) * 100:.2f}%"
120
+
121
+ # display the label and bounding box rectangle on the output
122
+ # frame
123
+ cv2.putText(
124
+ image,
125
+ label,
126
+ (startX, startY - 10),
127
+ cv2.FONT_HERSHEY_SIMPLEX,
128
+ 0.45,
129
+ color,
130
+ 2,
131
+ )
132
+ cv2.rectangle(image, (startX, startY), (endX, endY), color, 2)
133
+
134
+ # show the output image
135
+ cv2.imshow("Output", image)
136
+ cv2.waitKey(0)
137
+
138
+
139
+ def detect_mask_in_image(image, faceNet, maskNet):
140
+ # dimensions
141
+ (h, w) = image.shape[:2]
142
+
143
+ # construct a blob from the image
144
+ blob = cv2.dnn.blobFromImage(
145
+ image,
146
+ 1.0,
147
+ (300, 300),
148
+ (104.0, 177.0, 123.0),
149
+ ) # TODO: add to config
150
+
151
+ # pass the blob through the network and obtain the face detections
152
+ print("[INFO] computing face detections...")
153
+ faceNet.setInput(blob)
154
+ detections = faceNet.forward()
155
+
156
+ face_count = 0
157
+ # image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
158
+ # loop over the detections
159
+ for i in range(0, detections.shape[2]):
160
+ # extract the confidence associated with the detection
161
+ confidence = detections[0, 0, i, 2]
162
+ # print(f"[INFO] face {i}: {confidence}")
163
+ # filter out weak detections by ensuring the confidence is
164
+ # greater than the minimum confidence
165
+ if confidence > 0.5:
166
+ face_count += 1
167
+ # compute the (x, y)-coordinates of the object's bbox
168
+ box = detections[0, 0, i, 3:7] * np.array([w, h, w, h])
169
+ (startX, startY, endX, endY) = box.astype("int")
170
+
171
+ # ensure the bounding boxes fall within the dimensions of
172
+ # the frame
173
+ (startX, startY) = (max(0, startX), max(0, startY))
174
+ (endX, endY) = (min(w - 1, endX), min(h - 1, endY))
175
+
176
+ # extract the face ROI
177
+ # ordering, resize it to 224x224, and preprocess it
178
+ face = image[startY:endY, startX:endX]
179
+ face = cv2.cvtColor(face, cv2.COLOR_BGR2RGB)
180
+ face = cv2.resize(face, (224, 224))
181
+ face = img_to_array(face)
182
+ face = preprocess_input(face)
183
+ face = np.expand_dims(face, axis=0)
184
+
185
+ # pass the face through the model to determine if the face
186
+ # has a mask or not
187
+ (mask, withoutMask) = maskNet.predict(face)[0]
188
+
189
+ # determine the class label and color we'll use to draw
190
+ # the bounding box and text
191
+ label = "Mask" if mask > withoutMask else "No Mask"
192
+ color = (0, 255, 0) if label == "Mask" else (255, 0, 0)
193
+
194
+ # include the probability in the label
195
+ label = f"{label}: {max(mask, withoutMask) * 100:.2f}%"
196
+
197
+ # display the label & bbox rectangle on the output frame
198
+ cv2.putText(
199
+ image,
200
+ label,
201
+ (startX, startY - 10),
202
+ cv2.FONT_HERSHEY_SIMPLEX,
203
+ 0.45,
204
+ color,
205
+ 2,
206
+ )
207
+ cv2.rectangle(image, (startX, startY), (endX, endY), color, 2)
208
+ else:
209
+ break
210
+ text = f"[INFO] Detect {face_count} face(s)."
211
+ print(text)
212
+ cv2.putText(
213
+ image,
214
+ text,
215
+ (10, 30),
216
+ cv2.FONT_HERSHEY_SIMPLEX,
217
+ 0.70,
218
+ (0, 255, 0),
219
+ 2,
220
+ )
221
+ return image
222
+
223
+
224
+ if __name__ == "__main__":
225
+ mask_image()
src/image_processing.py ADDED
@@ -0,0 +1,60 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import cv2
2
+ import numpy as np
3
+ from itertools import groupby
4
+
5
+
6
+ def process_image(image, recognition_input_layer):
7
+ # Text detection models expect an image in grayscale format.
8
+ # IMPORTANT! This model enables reading only one line at time.
9
+ #image = cv2.cvtColor(image, cv2.COLOR_RGB2GRAY)
10
+
11
+
12
+ # Fetch the shape.
13
+ image_height, _ = image.shape
14
+
15
+ # B,C,H,W = batch size, number of channels, height, width.
16
+ _, _, H, W = recognition_input_layer.shape
17
+
18
+ # Calculate scale ratio between the input shape height and image height to resize the image.
19
+ scale_ratio = H / image_height
20
+
21
+ # Resize the image to expected input sizes.
22
+ resized_image = cv2.resize(
23
+ image, None, fx=scale_ratio, fy=scale_ratio, interpolation=cv2.INTER_AREA
24
+ )
25
+
26
+ # Pad the image to match input size, without changing aspect ratio.
27
+ resized_image = np.pad(
28
+ resized_image, ((0, 0), (0, W - resized_image.shape[1])), mode="edge"
29
+ )
30
+
31
+ # Reshape to network input shape.
32
+ input_image = resized_image[None, None, :, :]
33
+
34
+ return input_image
35
+
36
+
37
+ def recognize(image, compiled_model, recognition_input_layer, recognition_output_layer, letters):
38
+ input_image = process_image(image, recognition_input_layer)
39
+ # Run inference on the model
40
+ predictions = compiled_model([input_image])[recognition_output_layer]
41
+
42
+ # Remove a batch dimension.
43
+ predictions = np.squeeze(predictions)
44
+
45
+ # Run the `argmax` function to pick the symbols with the highest probability.
46
+ predictions_indexes = np.argmax(predictions, axis=1)
47
+
48
+ # Use the `groupby` function to remove concurrent letters, as required by CTC greedy decoding.
49
+ output_text_indexes = list(groupby(predictions_indexes))
50
+
51
+ # Remove grouper objects.
52
+ output_text_indexes, _ = np.transpose(output_text_indexes, (1, 0))
53
+
54
+ # Remove blank symbols.
55
+ output_text_indexes = output_text_indexes[output_text_indexes != 0]
56
+
57
+ # Assign letters to indexes from the output array.
58
+ output_text = [letters[letter_index] for letter_index in output_text_indexes]
59
+
60
+ return output_text
src/model2onnx.py ADDED
@@ -0,0 +1,41 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # import the necessary packages
2
+ import argparse
3
+
4
+ import onnx
5
+ import tf2onnx
6
+ from tensorflow.keras.models import load_model
7
+
8
+
9
+ def model2onnx():
10
+ # construct the argument parser and parse the arguments
11
+ ap = argparse.ArgumentParser()
12
+ ap.add_argument(
13
+ "-m",
14
+ "--model",
15
+ type=str,
16
+ default="mask_detector.model",
17
+ help="path to trained face mask detector model",
18
+ )
19
+ ap.add_argument(
20
+ "-o",
21
+ "--output",
22
+ type=str,
23
+ default="mask_detector.onnx",
24
+ help="path to trained face mask detector model",
25
+ )
26
+
27
+ args = vars(ap.parse_args())
28
+
29
+ # load the face mask detector model from disk
30
+ print("[INFO] loading face mask detector model...")
31
+ model = load_model(args["model"])
32
+ onnx_model, _ = tf2onnx.convert.from_keras(model, opset=13)
33
+
34
+ onnx_model.graph.input[0].type.tensor_type.shape.dim[0].dim_param = "?"
35
+ onnx_model.graph.output[0].type.tensor_type.shape.dim[0].dim_param = "?"
36
+
37
+ onnx.save(onnx_model, args["output"])
38
+
39
+
40
+ if __name__ == "__main__":
41
+ model2onnx()
src/search.py ADDED
@@ -0,0 +1,94 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import argparse
2
+ import os
3
+
4
+ import cv2
5
+ import requests
6
+ from requests import exceptions
7
+
8
+ ap = argparse.ArgumentParser()
9
+ ap.add_argument(
10
+ "-q",
11
+ "--query",
12
+ required=True,
13
+ help="search query to search Bing Image API for",
14
+ )
15
+ ap.add_argument(
16
+ "-o",
17
+ "--output",
18
+ required=True,
19
+ help="path to output directory of images",
20
+ )
21
+ args = vars(ap.parse_args())
22
+ API_KEY = "d8982f9e69a4437fa6e10715d1ed691d"
23
+ MAX_RESULTS = 500
24
+ GROUP_SIZE = 50
25
+ URL = "https://api.cognitive.microsoft.com/bing/v7.0/images/search"
26
+ EXCEPTIONS = {
27
+ IOError,
28
+ FileNotFoundError,
29
+ exceptions.RequestException,
30
+ exceptions.HTTPError,
31
+ exceptions.ConnectionError,
32
+ exceptions.Timeout,
33
+ }
34
+ term = args["query"]
35
+ headers = {"Ocp-Apim-Subscription-Key": API_KEY}
36
+ params = {"q": term, "offset": 0, "count": GROUP_SIZE}
37
+ print(f"[INFO] searching Bing API for '{term}'")
38
+ search = requests.get(URL, headers=headers, params=params)
39
+ search.raise_for_status()
40
+ results = search.json()
41
+ estNumResults = min(results["totalEstimatedMatches"], MAX_RESULTS)
42
+ print(
43
+ "[INFO] {} total results for '{}'".format(
44
+ estNumResults,
45
+ term,
46
+ ),
47
+ )
48
+ total = 0
49
+ for offset in range(0, estNumResults, GROUP_SIZE):
50
+ print(
51
+ "[INFO] making request for group {}-{} of {}...".format(
52
+ offset,
53
+ offset + GROUP_SIZE,
54
+ estNumResults,
55
+ ),
56
+ )
57
+ params["offset"] = offset
58
+ search = requests.get(URL, headers=headers, params=params)
59
+ search.raise_for_status()
60
+ results = search.json()
61
+ print(
62
+ "[INFO] saving images for group {}-{} of {}...".format(
63
+ offset,
64
+ offset + GROUP_SIZE,
65
+ estNumResults,
66
+ ),
67
+ )
68
+ for v in results["value"]:
69
+ try:
70
+ print("[INFO] fetching: {}".format(v["contentUrl"]))
71
+ r = requests.get(v["contentUrl"], timeout=30)
72
+ ext = v["contentUrl"][v["contentUrl"].rfind(".") :]
73
+ p = os.path.sep.join(
74
+ [
75
+ args["output"],
76
+ "{}{}".format(
77
+ str(total).zfill(8),
78
+ ext,
79
+ ),
80
+ ],
81
+ )
82
+ f = open(p, "wb")
83
+ f.write(r.content)
84
+ f.close()
85
+ except Exception as e:
86
+ if type(e) in EXCEPTIONS:
87
+ print("[INFO] skipping: {}".format(v["contentUrl"]))
88
+ continue
89
+ image = cv2.imread(p)
90
+ if image is None:
91
+ print(f"[INFO] deleting: {p}")
92
+ os.remove(p)
93
+ continue
94
+ total += 1
src/train_mask_detector.py ADDED
@@ -0,0 +1,192 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # USAGE
2
+ # python train_mask_detector.py --dataset dataset
3
+
4
+ import argparse
5
+ import os
6
+
7
+ import matplotlib.pyplot as plt
8
+ import numpy as np
9
+ from imutils import paths
10
+ from sklearn.metrics import classification_report
11
+ from sklearn.model_selection import train_test_split
12
+ from sklearn.preprocessing import LabelBinarizer
13
+ from tensorflow.keras.applications import MobileNetV2
14
+ from tensorflow.keras.applications.mobilenet_v2 import preprocess_input
15
+ from tensorflow.keras.layers import (
16
+ AveragePooling2D,
17
+ Dense,
18
+ Dropout,
19
+ Flatten,
20
+ Input,
21
+ )
22
+ from tensorflow.keras.models import Model
23
+ from tensorflow.keras.optimizers import Adam
24
+
25
+ # import the necessary packages
26
+ from tensorflow.keras.preprocessing.image import (
27
+ ImageDataGenerator,
28
+ img_to_array,
29
+ load_img,
30
+ )
31
+ from tensorflow.keras.utils import to_categorical
32
+
33
+ # construct the argument parser and parse the arguments
34
+ ap = argparse.ArgumentParser()
35
+ ap.add_argument(
36
+ "-d",
37
+ "--dataset",
38
+ required=True,
39
+ help="path to input dataset",
40
+ )
41
+ ap.add_argument(
42
+ "-p",
43
+ "--plot",
44
+ type=str,
45
+ default="plot.png",
46
+ help="path to output loss/accuracy plot",
47
+ )
48
+ ap.add_argument(
49
+ "-m",
50
+ "--model",
51
+ type=str,
52
+ default="mask_detector.model",
53
+ help="path to output face mask detector model",
54
+ )
55
+ args = vars(ap.parse_args())
56
+
57
+ # initialize the initial learning rate, number of epochs to train for,
58
+ # and batch size
59
+ INIT_LR = 1e-4
60
+ EPOCHS = 20
61
+ BS = 32
62
+
63
+ # grab the list of images in our dataset directory, then initialize
64
+ # the list of data (i.e., images) and class images
65
+ print("[INFO] loading images...")
66
+ imagePaths = list(paths.list_images(args["dataset"]))
67
+ data = []
68
+ labels = []
69
+
70
+ # loop over the image paths
71
+ for imagePath in imagePaths:
72
+ # extract the class label from the filename
73
+ label = imagePath.split(os.path.sep)[-2]
74
+
75
+ # load the input image (224x224) and preprocess it
76
+ image = load_img(imagePath, target_size=(224, 224))
77
+ image = img_to_array(image)
78
+ image = preprocess_input(image)
79
+
80
+ # update the data and labels lists, respectively
81
+ data.append(image)
82
+ labels.append(label)
83
+
84
+ # convert the data and labels to NumPy arrays
85
+ data = np.array(data, dtype="float32")
86
+ labels = np.array(labels)
87
+
88
+ # perform one-hot encoding on the labels
89
+ lb = LabelBinarizer()
90
+ labels = lb.fit_transform(labels)
91
+ labels = to_categorical(labels)
92
+
93
+ # partition the data into training and testing splits using 75% of
94
+ # the data for training and the remaining 25% for testing
95
+ (trainX, testX, trainY, testY) = train_test_split(
96
+ data,
97
+ labels,
98
+ test_size=0.20,
99
+ stratify=labels,
100
+ random_state=42,
101
+ )
102
+
103
+ # construct the training image generator for data augmentation
104
+ aug = ImageDataGenerator(
105
+ rotation_range=20,
106
+ zoom_range=0.15,
107
+ width_shift_range=0.2,
108
+ height_shift_range=0.2,
109
+ shear_range=0.15,
110
+ horizontal_flip=True,
111
+ fill_mode="nearest",
112
+ )
113
+
114
+ # load the MobileNetV2 network, ensuring the head FC layer sets are
115
+ # left off
116
+ baseModel = MobileNetV2(
117
+ weights="imagenet",
118
+ include_top=False,
119
+ input_tensor=Input(shape=(224, 224, 3)),
120
+ )
121
+
122
+ # construct the head of the model that will be placed on top of the
123
+ # the base model
124
+ headModel = baseModel.output
125
+ headModel = AveragePooling2D(pool_size=(7, 7))(headModel)
126
+ headModel = Flatten(name="flatten")(headModel)
127
+ headModel = Dense(128, activation="relu")(headModel)
128
+ headModel = Dropout(0.5)(headModel)
129
+ headModel = Dense(2, activation="softmax")(headModel)
130
+
131
+ # place the head FC model on top of the base model (this will become
132
+ # the actual model we will train)
133
+ model = Model(inputs=baseModel.input, outputs=headModel)
134
+
135
+ # loop over all layers in the base model and freeze them so they will
136
+ # *not* be updated during the first training process
137
+ for layer in baseModel.layers:
138
+ layer.trainable = False
139
+
140
+ # compile our model
141
+ print("[INFO] compiling model...")
142
+ opt = Adam(lr=INIT_LR, decay=INIT_LR / EPOCHS)
143
+ model.compile(
144
+ loss="binary_crossentropy",
145
+ optimizer=opt,
146
+ metrics=["accuracy"],
147
+ )
148
+
149
+ # train the head of the network
150
+ print("[INFO] training head...")
151
+ H = model.fit(
152
+ aug.flow(trainX, trainY, batch_size=BS),
153
+ steps_per_epoch=len(trainX) // BS,
154
+ validation_data=(testX, testY),
155
+ validation_steps=len(testX) // BS,
156
+ epochs=EPOCHS,
157
+ )
158
+
159
+ # make predictions on the testing set
160
+ print("[INFO] evaluating network...")
161
+ predIdxs = model.predict(testX, batch_size=BS)
162
+
163
+ # for each image in the testing set we need to find the index of the
164
+ # label with corresponding largest predicted probability
165
+ predIdxs = np.argmax(predIdxs, axis=1)
166
+
167
+ # show a nicely formatted classification report
168
+ print(
169
+ classification_report(
170
+ testY.argmax(axis=1),
171
+ predIdxs,
172
+ target_names=lb.classes_,
173
+ ),
174
+ )
175
+
176
+ # serialize the model to disk
177
+ print("[INFO] saving mask detector model...")
178
+ model.save(args["model"], save_format="h5")
179
+
180
+ # plot the training loss and accuracy
181
+ N = EPOCHS
182
+ plt.style.use("ggplot")
183
+ plt.figure()
184
+ plt.plot(np.arange(0, N), H.history["loss"], label="train_loss")
185
+ plt.plot(np.arange(0, N), H.history["val_loss"], label="val_loss")
186
+ plt.plot(np.arange(0, N), H.history["accuracy"], label="train_acc")
187
+ plt.plot(np.arange(0, N), H.history["val_accuracy"], label="val_acc")
188
+ plt.title("Training Loss and Accuracy")
189
+ plt.xlabel("Epoch #")
190
+ plt.ylabel("Loss/Accuracy")
191
+ plt.legend(loc="lower left")
192
+ plt.savefig(args["plot"])