init

.gitignore

@@ -0,0 +1,134 @@
+# ignore map, miou, datasets
+datasets/archive/Dataset
+model_data/best_epoch_weights.pth
+metrics_out/
+
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+pip-wheel-metadata/
+share/python-wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# PyInstaller
+#  Usually these files are written by a python script from a template
+#  before PyInstaller builds the exe, so as to inject date/other infos into it.
+*.manifest
+*.spec
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+*.py,cover
+.hypothesis/
+.pytest_cache/
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+local_settings.py
+db.sqlite3
+db.sqlite3-journal
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# IPython
+profile_default/
+ipython_config.py
+
+# pyenv
+.python-version
+
+# pipenv
+#   According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
+#   However, in case of collaboration, if having platform-specific dependencies or dependencies
+#   having no cross-platform support, pipenv may install dependencies that don't work, or not
+#   install all needed dependencies.
+#Pipfile.lock
+
+# PEP 582; used by e.g. github.com/David-OConnor/pyflow
+__pypackages__/
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+.dmypy.json
+dmypy.json
+
+# Pyre type checker
+.pyre/

AutoML.py

@@ -0,0 +1,102 @@
+'''
+Author: [egrt]
+Date: 2022-08-14 09:37:12
+LastEditors: [egrt]
+LastEditTime: 2022-08-17 20:34:36
+Description: 
+'''
+import numpy as np
+import pandas as pd
+import pickle
+from sklearn.preprocessing import LabelEncoder
+
+def show_config(**kwargs):
+    print('Configurations:')
+    print('-' * 70)
+    print('|%25s | %40s|' % ('keys', 'values'))
+    print('-' * 70)
+    for key, value in kwargs.items():
+        print('|%25s | %40s|' % (str(key), str(value)))
+    print('-' * 70)
+    
+#--------------------------------------------#
+#   使用自己训练好的模型预测需要修改3个参数
+#   model_path和classes_path和backbone都需要修改！
+#--------------------------------------------#
+class Classification(object):
+    _defaults = {
+        #--------------------------------------------------------------------------#
+        #   使用自己训练好的模型进行预测一定要修改model_path和classes_path！
+        #   model_path指向logs文件夹下的权值文件，classes_path指向model_data下的txt
+        #   如果出现shape不匹配，同时要注意训练时的model_path和classes_path参数的修改
+        #--------------------------------------------------------------------------#
+        "model_path"        : 'model_data/automl_v2.pkl',
+        "train_path"        : 'datasets/archive/artworks.csv',
+        #-------------------------------#
+        #   是否使用Cuda
+        #   没有GPU可以设置成False
+        #-------------------------------#
+        "cuda"              : False
+    }
+
+    @classmethod
+    def get_defaults(cls, n):
+        if n in cls._defaults:
+            return cls._defaults[n]
+        else:
+            return "Unrecognized attribute name '" + n + "'"
+
+    #---------------------------------------------------#
+    #   初始化classification
+    #---------------------------------------------------#
+    def __init__(self, **kwargs):
+        self.__dict__.update(self._defaults)
+        for name, value in kwargs.items():
+            setattr(self, name, value)
+
+        #---------------------------------------------------#
+        #   获得种类
+        #---------------------------------------------------#
+        self.num_classes = 1
+        self.train_data  = pd.read_csv(self.train_path)
+        self.generate()
+        
+        show_config(**self._defaults)
+
+    #---------------------------------------------------#
+    #   获得所有的分类
+    #---------------------------------------------------#
+    def generate(self):
+        #---------------------------------------------------#
+        #   载入模型与权值
+        #---------------------------------------------------#
+        with open("model_data/automl_v2.pkl", "rb") as f:
+            self.automl = pickle.load(f)
+
+    def detect_one(self, name, date, level, classification, height, width):
+        # 读取数据集
+        train_data   = self.train_data
+        ArtistID     = train_data.loc[train_data["Name"] == name, "Artist ID"][0]
+        # 对输入数据进行编码
+        la_Catalogue = LabelEncoder()
+        la_Catalogue.fit(train_data["Catalogue"])
+        Catalogue = la_Catalogue.transform(["Y"])
+
+        la_Department = LabelEncoder()
+        la_Department.fit(train_data["Department"])
+        Department = la_Department.transform([level])
+
+        la_Classification = LabelEncoder()
+        la_Classification.fit(train_data["Classification"])
+        Classification = la_Classification.transform([classification])
+        test_dict = {'Artist ID':ArtistID,'Date':date,'Department':Department,'Classification':Classification,
+            "Height (cm)":height, "Width (cm)":width}
+        test_data = pd.DataFrame(test_dict)
+
+        pred = self.automl.predict(test_data)
+        return int(pred[0])
+
+if __name__ == "__main__":
+    classfication = Classification()
+    # classfication.get_result()
+    classfication.detect_one("陈冠夫", 1975, '国家级', '中国山水画', 50, 50)

app.py

@@ -0,0 +1,55 @@
+'''
+Author: [egrt]
+Date: 2022-08-05 15:30:04
+LastEditors: [egrt]
+LastEditTime: 2022-08-17 20:36:44
+Description: 
+'''
+import os
+os.system('pip install requirements.txt')
+import gradio as gr
+from AutoML import Classification
+
+classfication = Classification()
+
+with gr.Blocks() as demo:
+    with gr.Row():
+        with gr.Column():
+            name_box    = gr.Textbox(label="艺术家姓名")
+            date_box    = gr.Number(label="艺术品创作时间(年)")
+            classification_box = gr.CheckboxGroup(label="艺术品类型", choices=["中国山水画", "中国花鸟画", "风景油画",
+                "水粉画", "油画", "布面油画", "没骨画", "中国画"])
+            level_box   = gr.CheckboxGroup(label="艺术家级别", choices=["国家级", "省级"])
+            height_box  = gr.Number(label="艺术品高度(cm)")
+            width_box   = gr.Number(label="艺术品宽度(cm)")
+            
+        with gr.Column():
+            title_box   = gr.Markdown("## 艺术品价值预测Demo")
+            image_box   = gr.Image(label="艺术品照片", type='pil')
+            with gr.Row():
+                submit_btn = gr.Button("开始预测")
+            error_box   = gr.Textbox(label="错误", visible=False)
+            with gr.Row(visible=False) as output_col:
+                prices_box = gr.Number(label="预测的艺术品价格(￥)")
+                
+
+    def submit(name, date, classification, level, height, width, image):
+        # 填写信息错误则显示报错提示框
+        if len(name)==0 or len(classification)==0 or len(level) == 0 or image==None:
+            return {error_box: gr.update(value="请填写所有信息", visible=True)}
+        else:
+            # 正确预测之后则去除报错提示框
+            error_box: gr.update(visible=False)
+            prices = classfication.detect_one(name, date, level[0], classification[0], height, width)
+            return {
+                output_col: gr.update(visible=True),
+                prices_box: int(prices),
+            }
+
+    submit_btn.click(
+        submit,
+        [name_box, date_box, classification_box, level_box, height_box, width_box, image_box],
+        [error_box, prices_box, output_col],
+    )
+
+demo.launch()

datasets/archive/artists.csv

@@ -0,0 +1,36 @@
+Artist ID,Name,Nationality,Gender,Birth Year,Death Year
+1,Ai Xuan 艾軒,Zhejiang,Male,1947,
+2,Chang DaiChien 張大千,Neijiang,Male,1899,1983
+3,Chu Chu 儲楚,Hangzhou,Female,1975,
+4,Giuseppe Castiglione 郎世宁,Italy,Male,1688,1766
+5,Gu Wenda 谷文达,Shanghai,Male,1955,
+6,He Baili 何百里,GuangZhou,Male,1945,
+7,He Jilan 何紀嵐,Hong Kong,Male,1974,
+9,Irene Chou 周綠雲,Shanghai,Female,1924,2011
+10,Jin TingBiao 金廷标,Huzhou,Male,1767,
+11,Li FangYing 李方膺,Jiangsu,Male,1696,1755
+12,Li Zijian 李自健,ShaoYang,Male,1954,
+13,Ling Jian 凌健,ShanDong,Male,1963,
+15,Liu KuoSung 刘国松,AnHui,Male,1932,
+16,Liu YuanShou 劉元壽,Beijing,Male,1967,
+18,Lu XinJian 陆新建,Jiangsu,Male,1977,
+19,Luo ZhongLi 羅中立,ChongQing,Male,1948,
+21,Qin Feng 秦风,XinJiang,Male,1961,
+22,Qiu Ying 仇英,Jiangsu,Male,1494,1552
+23,Wai Ming,Canton,Male,1938,
+24,Wang Hui 王翚,SuZhou,Male,1632,1717
+25,Yao WenHan 姚文瀚,顺天（今北京）,Male,公元18世纪,公元18世纪
+26,Yu ZhiDing 禹之鼎,YangZhou,Male,1647,1709
+27,Zhang XiaoGang 张晓刚,KunMing,Male,1958,
+28,Zhao KaiLin 赵开霖,BengBu,Male,1961,
+30,Zou ChuanAn 邹传安,XinHua,Male,1941,
+31,,,,,
+32,,,,,
+33,,,,,
+34,,,,,
+35,,,,,
+36,,,,,
+37,,,,,
+38,,,,,
+39,,,,,
+40,,,,,

datasets/archive/artworks.csv

@@ -0,0 +1,319 @@
+Artwork ID,Title,Artist ID,Name,Date,Medium,Dimensions,Catalogue,Department,Classification,Height (cm),Length (cm),Width (cm),Depth (cm),Weight (kg),Duration (s),Prices
+1,拟古重彩山水之一,1,陈冠夫,1975,,,Y,国家级,中国山水画,50,,50,,,陈冠夫/Chen_Guang_Fu_01.jpg,4000
+2,拟古重彩山水之二,1,陈冠夫,1975,,,Y,国家级,中国山水画,50,,50,,,陈冠夫/Chen_Guang_Fu_02.jpg,4000
+3,拟古重彩山水之三,1,陈冠夫,1975,,,Y,国家级,中国山水画,50,,50,,,陈冠夫/Chen_Guang_Fu_03.jpg,4000
+4,团扇系列之一,1,陈冠夫,1975,,,Y,国家级,中国山水画,45,,45,,,陈冠夫/Chen_Guang_Fu_04.jpg,4000
+5,团扇系列之二,1,陈冠夫,1975,,,Y,国家级,中国山水画,45,,45,,,陈冠夫/Chen_Guang_Fu_05.jpg,4000
+6,团扇系列之三,1,陈冠夫,1975,,,Y,国家级,中国山水画,45,,45,,,陈冠夫/Chen_Guang_Fu_06.jpg,4000
+7,团扇系列之四,1,陈冠夫,1975,,,Y,国家级,中国山水画,45,,45,,,陈冠夫/Chen_Guang_Fu_07.jpg,4000
+8,团扇系列之五,1,陈冠夫,1975,,,Y,国家级,中国山水画,45,,45,,,陈冠夫/Chen_Guang_Fu_08.jpg,4000
+9,团扇系列之六,1,陈冠夫,1975,,,Y,国家级,中国山水画,45,,45,,,陈冠夫/Chen_Guang_Fu_09.jpg,4000
+10,天地有大美而不言,1,陈冠夫,1975,,,Y,国家级,中国山水画,67,,67,,,陈冠夫/Chen_Guang_Fu_10.jpg,4000
+11,春山淡妆,1,陈冠夫,1975,,,Y,国家级,中国山水画,60,,50,,,陈冠夫/Chen_Guang_Fu_11.jpg,4000
+12,意象.蜀山苍秀,1,陈冠夫,1975,,,Y,国家级,中国山水画,220,,160,,,陈冠夫/Chen_Guang_Fu_12.jpg,4000
+13,蜀山隐居,1,陈冠夫,1975,,,Y,国家级,中国山水画,55,,35,,,陈冠夫/Chen_Guang_Fu_13.jpg,4000
+14,山居问道,1,陈冠夫,1975,,,Y,国家级,中国山水画,55,,35,,,陈冠夫/Chen_Guang_Fu_14.jpg,4000
+15,春山悟道,1,陈冠夫,1975,,,Y,国家级,中国山水画,70,,40,,,陈冠夫/Chen_Guang_Fu_15.jpg,4000
+16,荆楚风雨过后 依旧山明水秀,1,陈冠夫,1975,,,Y,国家级,中国山水画,50,,120,,,陈冠夫/Chen_Guang_Fu_16.jpg,4000
+17,松涛.意象,1,陈冠夫,1975,,,Y,国家级,中国山水画,170,,90,,,陈冠夫/Chen_Guang_Fu_17.jpg,4000
+18,意象.蜀山苍秀,1,陈冠夫,1975,,,Y,国家级,中国山水画,220,,160,,,陈冠夫/Chen_Guang_Fu_18.jpg,4000
+19,天地·镜像系列之二,1,陈冠夫,1975,,,Y,国家级,中国山水画,45,,180,,,陈冠夫/Chen_Guang_Fu_19.jpg,4000
+20,意象?·水墨系列之三,1,陈冠夫,1975,,,Y,国家级,中国山水画,147,,50,,,陈冠夫/Chen_Guang_Fu_20.jpg,4000
+21,蜀山烟岚,1,陈冠夫,1975,,,Y,国家级,中国山水画,108,,45,,,陈冠夫/Chen_Guang_Fu_21.jpg,4000
+22,意象.水云间之一,1,陈冠夫,1975,,,Y,国家级,中国山水画,35,,67,,,陈冠夫/Chen_Guang_Fu_22.jpg,4000
+23,意象.水云间山水之二,1,陈冠夫,1975,,,Y,国家级,中国山水画,35,,67,,,陈冠夫/Chen_Guang_Fu_23.jpg,4000
+24,秋山流云,1,陈冠夫,1975,,,Y,国家级,中国山水画,45,,180,,,陈冠夫/Chen_Guang_Fu_24.jpg,4000
+25,墨彩系列之一,1,陈冠夫,1975,,,Y,国家级,中国山水画,45,,70,,,陈冠夫/Chen_Guang_Fu_25.jpg,4000
+26,春晓,2,梁时民,,,,Y,国家级,中国花鸟画,50,,50,,,梁时民/Liang_Shi_Ming_01.jpg,67500
+27,荷塘清露,2,梁时民,,,,Y,国家级,中国花鸟画,68,,68,,,梁时民/Liang_Shi_Ming_02.jpg,124848
+28,鸳鸯无限意 任尔烟鱼游,2,梁时民,,,,Y,国家级,中国花鸟画,136,,68,,,梁时民/Liang_Shi_Ming_03.jpg,249696
+29,百叶莲花万重香,2,梁时民,,,,Y,国家级,中国花鸟画,50,,50,,,梁时民/Liang_Shi_Ming_04.jpg,67500
+30,渊冰厚三尺 素雪覆千里,2,梁时民,,,,Y,国家级,中国花鸟画,68,,135,,,梁时民/Liang_Shi_Ming_05.jpg,247860
+31,晚秋,2,梁时民,,,,Y,国家级,中国花鸟画,136,,68,,,梁时民/Liang_Shi_Ming_06.jpg,249696
+32,风摇青玉枝,2,梁时民,,,,Y,国家级,中国花鸟画,68,,68,,,梁时民/Liang_Shi_Ming_07.jpg,124848
+33,暖冬,2,梁时民,,,,Y,国家级,中国花鸟画,192,,181,,,梁时民/Liang_Shi_Ming_08.jpg,938304
+34,初雪,2,梁时民,,,,Y,国家级,中国花鸟画,35,,35,,,梁时民/Liang_Shi_Ming_09.jpg,33075
+35,素秋,2,梁时民,,,,Y,国家级,中国花鸟画,68,,40,,,梁时民/Liang_Shi_Ming_10.jpg,73440
+36,包谷林,2,梁时民,,,,Y,国家级,中国花鸟画,181,,192,,,梁时民/Liang_Shi_Ming_11.jpg,938304
+37,山里回忆,3,邵智,,,,Y,国家级,风景油画,60,,50,,,邵智/Shao_Zhi_01.jpg,10000
+38,车里子,3,邵智,,,,Y,国家级,风景油画,60,,50,,,邵智/Shao_Zhi_02.jpg,10000
+39,硕果,3,邵智,,,,Y,国家级,风景油画,50,,60,,,邵智/Shao_Zhi_03.jpg,10000
+40,多仔多福,3,邵智,,,,Y,国家级,风景油画,50,,50,,,邵智/Shao_Zhi_04.jpg,10000
+41,一篮秋歌,3,邵智,,,,Y,国家级,风景油画,50,,50,,,邵智/Shao_Zhi_05.jpg,10000
+42,秋韵,3,邵智,,,,Y,国家级,风景油画,50,,50,,,邵智/Shao_Zhi_06.jpg,10000
+43,绿萝,3,邵智,,,,Y,国家级,风景油画,100,,60,,,邵智/Shao_Zhi_07.jpg,10000
+44,电吹风,3,邵智,,,,Y,国家级,风景油画,100,,60,,,邵智/Shao_Zhi_08.jpg,10000
+45,岁月如歌,3,邵智,,,,Y,国家级,风景油画,65,,85,,,邵智/Shao_Zhi_09.jpg,10000
+46,寻梅,3,邵智,,,,Y,国家级,风景油画,85,,65,,,邵智/Shao_Zhi_10.jpg,10000
+47,清影,3,邵智,,,,Y,国家级,风景油画,60,,50,,,邵智/Shao_Zhi_11.jpg,10000
+48,荷香幽蓝,3,邵智,,,,Y,国家级,风景油画,50,,60,,,邵智/Shao_Zhi_12.jpg,10000
+49,今夕何夕,3,邵智,,,,Y,国家级,风景油画,60,,50,,,邵智/Shao_Zhi_13.jpg,10000
+50,笔下春秋,3,邵智,,,,Y,国家级,风景油画,60,,50,,,邵智/Shao_Zhi_14.jpg,10000
+51,西游记,3,邵智,,,,Y,国家级,风景油画,50,,50,,,邵智/Shao_Zhi_15.jpg,10000
+52,爷爷的灯,3,邵智,,,,Y,国家级,风景油画,90,,120,,,邵智/Shao_Zhi_16.jpg,10000
+53,婚事,3,邵智,,,,Y,国家级,风景油画,80,,110,,,邵智/Shao_Zhi_17.jpg,10000
+54,兄弟,3,邵智,,,,Y,国家级,风景油画,80,,110,,,邵智/Shao_Zhi_18.jpg,10000
+55,解放大西南,3,邵智,,,,Y,国家级,风景油画,160,,220,,,邵智/Shao_Zhi_19.jpg,10000
+56,昊天,3,邵智,,,,Y,国家级,风景油画,110,,90,,,邵智/Shao_Zhi_20.jpg,10000
+57,回娘家,3,邵智,,,,Y,国家级,风景油画,80,,100,,,邵智/Shao_Zhi_21.jpg,10000
+58,回望家园,3,邵智,,,,Y,国家级,风景油画,110,,90,,,邵智/Shao_Zhi_22.jpg,10000
+59,古镇晨曲,3,邵智,,,,Y,国家级,风景油画,70,,100,,,邵智/Shao_Zhi_23.jpg,10000
+60,故乡的小河2,3,邵智,,,,Y,国家级,风景油画,65,,75,,,邵智/Shao_Zhi_24.jpg,10000
+61,彝海情深,3,邵智,,,,Y,国家级,风景油画,70,,120,,,邵智/Shao_Zhi_25.jpg,10000
+62,最后的净土,3,邵智,,,,Y,国家级,风景油画,120,,360,,,邵智/Shao_Zhi_26.jpg,10000
+63,吉祥富贵,4,谭昌镕,1933,,,Y,国家级,中国花鸟画,136,,68,,,谭昌镕/Tang_Chang_Rong_01.jpg,249696
+64,吉祥富贵图,4,谭昌镕,1933,,,Y,国家级,中国花鸟画,136,,68,,,谭昌镕/Tang_Chang_Rong_02.jpg,249696
+65,春来了,4,谭昌镕,1933,,,Y,国家级,中国花鸟画,136,,68,,,谭昌镕/Tang_Chang_Rong_03.jpg,249696
+66,迎春,4,谭昌镕,1933,,,Y,国家级,中国花鸟画,136,,68,,,谭昌镕/Tang_Chang_Rong_04.jpg,249696
+67,飞吧,4,谭昌镕,1933,,,Y,国家级,中国花鸟画,136,,68,,,谭昌镕/Tang_Chang_Rong_05.jpg,249696
+68,无题,4,谭昌镕,1933,,,Y,国家级,中国花鸟画,136,,68,,,谭昌镕/Tang_Chang_Rong_06.jpg,249696
+69,春长在,4,谭昌镕,1933,,,Y,国家级,中国花鸟画,136,,68,,,谭昌镕/Tang_Chang_Rong_07.jpg,249696
+70,无 题,4,谭昌镕,1933,,,Y,国家级,中国花鸟画,136,,68,,,谭昌镕/Tang_Chang_Rong_08.jpg,249696
+71,初温,4,谭昌镕,1933,,,Y,国家级,中国花鸟画,136,,68,,,谭昌镕/Tang_Chang_Rong_09.jpg,249696
+72,无 题,4,谭昌镕,1933,,,Y,国家级,中国花鸟画,136,,68,,,谭昌镕/Tang_Chang_Rong_10.jpg,249696
+73,腾 飞,4,谭昌镕,1933,,,Y,国家级,中国花鸟画,136,,68,,,谭昌镕/Tang_Chang_Rong_11.jpg,249696
+74,秋趣,4,谭昌镕,1933,,,Y,国家级,中国花鸟画,136,,68,,,谭昌镕/Tang_Chang_Rong_12.jpg,249696
+75,初温,4,谭昌镕,1933,,,Y,国家级,中国花鸟画,136,,68,,,谭昌镕/Tang_Chang_Rong_13.jpg,249696
+76,百花系列之牡丹,5,汪晓灵,,,,Y,国家级,中国花鸟画,69,,70,,,汪晓灵/Wang_Xiao_Ling_01.jpg,56511
+77,百花系列之白芍药,5,汪晓灵,,,,Y,国家级,中国花鸟画,69,,70,,,汪晓灵/Wang_Xiao_Ling_02.jpg,56511
+78,百花系列之高山杜鹃,5,汪晓灵,,,,Y,国家级,中国花鸟画,69,,70,,,汪晓灵/Wang_Xiao_Ling_03.jpg,56511
+79,百花系列之红山茶,5,汪晓灵,,,,Y,国家级,中国花鸟画,69,,70,,,汪晓灵/Wang_Xiao_Ling_04.jpg,56511
+80,百花系列之箭杆花,5,汪晓灵,,,,Y,国家级,中国花鸟画,69,,70,,,汪晓灵/Wang_Xiao_Ling_05.jpg,56511
+81,百花系列之红棉,5,汪晓灵,,,,Y,国家级,中国花鸟画,69,,70,,,汪晓灵/Wang_Xiao_Ling_06.jpg,56511
+82,百花系列之红掌,5,汪晓灵,,,,Y,国家级,中国花鸟画,69,,69,,,汪晓灵/Wang_Xiao_Ling_07.jpg,55703.7
+83,百花系列之牵牛花,5,汪晓灵,,,,Y,国家级,中国花鸟画,69,,70,,,汪晓灵/Wang_Xiao_Ling_08.jpg,56511
+84,百花系列之荷花,5,汪晓灵,,,,Y,国家级,中国花鸟画,69,,70,,,汪晓灵/Wang_Xiao_Ling_09.jpg,56511
+85,百花系列之梨花,5,汪晓灵,,,,Y,国家级,中国花鸟画,69,,69,,,汪晓灵/Wang_Xiao_Ling_10.jpg,55703.7
+86,百花系列之铁脚海棠,5,汪晓灵,,,,Y,国家级,中国花鸟画,69,,69,,,汪晓灵/Wang_Xiao_Ling_11.jpg,55703.7
+87,百花系列之旱金莲,5,汪晓灵,,,,Y,国家级,中国花鸟画,69,,69,,,汪晓灵/Wang_Xiao_Ling_12.jpg,55703.7
+88,百花系列之黄花槐,5,汪晓灵,,,,Y,国家级,中国花鸟画,69,,69,,,汪晓灵/Wang_Xiao_Ling_13.jpg,55703.7
+89,百花系列之菊花,5,汪晓灵,,,,Y,国家级,中国花鸟画,69,,69,,,汪晓灵/Wang_Xiao_Ling_14.jpg,55703.7
+90,秋晨,5,汪晓灵,,,,Y,国家级,中国花鸟画,34,,45,,,汪晓灵/Wang_Xiao_Ling_15.jpg,17901
+91,泸沽湖上,5,汪晓灵,,,,Y,国家级,中国花鸟画,89,,45.5,,,汪晓灵/Wang_Xiao_Ling_16.jpg,47379.15
+92,晨牧,5,汪晓灵,,,,Y,国家级,中国花鸟画,69,,69,,,汪晓灵/Wang_Xiao_Ling_17.jpg,55703.7
+93,荷塘月色,5,汪晓灵,,,,Y,国家级,中国花鸟画,69,,69,,,汪晓灵/Wang_Xiao_Ling_18.jpg,55703.7
+94,良宵,5,汪晓灵,,,,Y,国家级,中国花鸟画,69,,69,,,汪晓灵/Wang_Xiao_Ling_19.jpg,55703.7
+95,蓝天之歌,5,汪晓灵,,,,Y,国家级,中国花鸟画,35,,45,,,汪晓灵/Wang_Xiao_Ling_20.jpg,18427.5
+96,高山流水图,6,车晓葵,1953,,,Y,国家级,中国花鸟画,35,,35,,,车晓葵/Che_Xiao_Kui_01.jpg,5000
+97,好风光,6,车晓葵,1953,,,Y,国家级,中国花鸟画,35,,35,,,车晓葵/Che_Xiao_Kui_02.jpg,5000
+98,赞梅,6,车晓葵,1953,,,Y,国家级,中国花鸟画,35,,35,,,车晓葵/Che_Xiao_Kui_03.jpg,5000
+99,枇杷果果,6,车晓葵,1953,,,Y,国家级,中国花鸟画,35,,35,,,车晓葵/Che_Xiao_Kui_04.jpg,5000
+100,雅香,6,车晓葵,1953,,,Y,国家级,中国花鸟画,35,,35,,,车晓葵/Che_Xiao_Kui_05.jpg,5000
+101,情趣,6,车晓葵,1953,,,Y,国家级,中国花鸟画,35,,35,,,车晓葵/Che_Xiao_Kui_06.jpg,5000
+102,开口笑,6,车晓葵,1953,,,Y,国家级,中国花鸟画,35,,35,,,车晓葵/Che_Xiao_Kui_07.jpg,5000
+103,天香,6,车晓葵,1953,,,Y,国家级,中国花鸟画,69,,69,,,车晓葵/Che_Xiao_Kui_08.jpg,5000
+104,荷塘秋果,6,车晓葵,1953,,,Y,国家级,中国花鸟画,69,,69,,,车晓葵/Che_Xiao_Kui_09.jpg,5000
+105,紫藤小鸟图,6,车晓葵,1953,,,Y,国家级,中国花鸟画,69,,69,,,车晓葵/Che_Xiao_Kui_10.jpg,5000
+106,春花正茂,6,车晓葵,1953,,,Y,国家级,中国花鸟画,35,,70,,,车晓葵/Che_Xiao_Kui_11.jpg,5000
+107,国色天香,6,车晓葵,1953,,,Y,国家级,中国花鸟画,35,,110,,,车晓葵/Che_Xiao_Kui_12.jpg,5000
+108,荷塘月色,6,车晓葵,1953,,,Y,国家级,中国花鸟画,52,,155,,,车晓葵/Che_Xiao_Kui_13.jpg,5000
+109,春暖花开春常在,6,车晓葵,1953,,,Y,国家级,中国花鸟画,52,,223,,,车晓葵/Che_Xiao_Kui_14.jpg,5000
+110,春、夏、秋、冬,6,车晓葵,1953,,,Y,国家级,中国花鸟画,50,,208,,,车晓葵/Che_Xiao_Kui_15.jpg,5000
+111,春满人间,6,车晓葵,1953,,,Y,国家级,中国花鸟画,69,,138,,,车晓葵/Che_Xiao_Kui_16.jpg,5000
+112,紫气东来,6,车晓葵,1953,,,Y,国家级,中国花鸟画,69,,138,,,车晓葵/Che_Xiao_Kui_17.jpg,5000
+113,志在云天,6,车晓葵,1953,,,Y,国家级,中国花鸟画,69,,138,,,车晓葵/Che_Xiao_Kui_18.jpg,5000
+114,荷塘月色,6,车晓葵,1953,,,Y,国家级,中国花鸟画,69,,136,,,车晓葵/Che_Xiao_Kui_19.jpg,5000
+115,紫气迎春,6,车晓葵,1953,,,Y,国家级,中国花鸟画,120,,245,,,车晓葵/Che_Xiao_Kui_20.jpg,5000
+116,出水荷花香,6,车晓葵,1953,,,Y,国家级,中国花鸟画,97,,180,,,车晓葵/Che_Xiao_Kui_21.jpg,5000
+117,福禄大吉图,6,车晓葵,1953,,,Y,国家级,中国花鸟画,35,,140,,,车晓葵/Che_Xiao_Kui_22.jpg,5000
+118,思语,6,车晓葵,1953,,,Y,国家级,中国花鸟画,35,,120,,,车晓葵/Che_Xiao_Kui_23.jpg,5000
+119,凌云图,6,车晓葵,1953,,,Y,国家级,中国花鸟画,35,,140,,,车晓葵/Che_Xiao_Kui_24.jpg,5000
+117,,7,何学斌,,,,Y,国家级,中国山水画,120,,40,,,何学斌/He_Xue_Bin_01.jpg,216000
+118,,7,何学斌,,,,Y,国家级,中国山水画,120,,40,,,何学斌/He_Xue_Bin_02.jpg,216000
+119,,7,何学斌,,,,Y,国家级,中国山水画,120,,40,,,何学斌/He_Xue_Bin_03.jpg,216000
+120,,7,何学斌,,,,Y,国家级,中国山水画,120,,40,,,何学斌/He_Xue_Bin_04.jpg,216000
+121,,7,何学斌,,,,Y,国家级,中国山水画,120,,40,,,何学斌/He_Xue_Bin_05.jpg,216000
+122,,7,何学斌,,,,Y,国家级,中国山水画,98,,180,,,何学斌/He_Xue_Bin_06.jpg,793800
+123,,7,何学斌,,,,Y,国家级,中国山水画,120,,40,,,何学斌/He_Xue_Bin_07.jpg,216000
+124,,7,何学斌,,,,Y,国家级,中国山水画,120,,40,,,何学斌/He_Xue_Bin_08.jpg,216000
+125,,7,何学斌,,,,Y,国家级,中国山水画,120,,40,,,何学斌/He_Xue_Bin_09.jpg,216000
+126,,7,何学斌,,,,Y,国家级,中国山水画,120,,40,,,何学斌/He_Xue_Bin_10.jpg,216000
+127,,7,何学斌,,,,Y,国家级,中国山水画,120,,50,,,何学斌/He_Xue_Bin_11.jpg,270000
+128,,7,何学斌,,,,Y,国家级,中国山水画,120,,50,,,何学斌/He_Xue_Bin_12.jpg,270000
+129,,7,何学斌,,,,Y,国家级,中国山水画,120,,50,,,何学斌/He_Xue_Bin_13.jpg,216000
+130,,7,何学斌,,,,Y,国家级,中国山水画,120,,40,,,何学斌/He_Xue_Bin_14.jpg,216000
+131,九寨沟小景一,8,冷蓉,,,,Y,国家级,中国山水画,45,,45,,,冷蓉/Leng_Rong_01.jpg,5000
+132,九寨沟小景二,8,冷蓉,,,,Y,国家级,中国山水画,45,,45,,,冷蓉/Leng_Rong_02.jpg,5000
+133,九寨沟小景三,8,冷蓉,,,,Y,国家级,中国山水画,45,,45,,,冷蓉/Leng_Rong_03.jpg,5000
+134,小品一,8,冷蓉,,,,Y,国家级,中国山水画,45,,45,,,冷蓉/Leng_Rong_04.jpg,5000
+135,小品二,8,冷蓉,,,,Y,国家级,中国山水画,45,,45,,,冷蓉/Leng_Rong_05.jpg,5000
+136,小品三,8,冷蓉,,,,Y,国家级,中国山水画,45,,45,,,冷蓉/Leng_Rong_06.jpg,5000
+137,小品四,8,冷蓉,,,,Y,国家级,中国山水画,45,,45,,,冷蓉/Leng_Rong_07.jpg,5000
+138,小品五,8,冷蓉,,,,Y,国家级,中国山水画,45,,45,,,冷蓉/Leng_Rong_08.jpg,5000
+139,小品六,8,冷蓉,,,,Y,国家级,中国山水画,45,,45,,,冷蓉/Leng_Rong_09.jpg,5000
+140,黄龙溪写生一,8,冷蓉,,,,Y,国家级,中国山水画,50,,50,,,冷蓉/Leng_Rong_10.jpg,5000
+141,黄龙溪写生二,8,冷蓉,,,,Y,国家级,中国山水画,50,,50,,,冷蓉/Leng_Rong_11.jpg,5000
+142,黄龙溪写生三,8,冷蓉,,,,Y,国家级,中国山水画,50,,50,,,冷蓉/Leng_Rong_12.jpg,5000
+143,黄龙溪写生四,8,冷蓉,,,,Y,国家级,中国山水画,50,,50,,,冷蓉/Leng_Rong_13.jpg,5000
+144,达古山下,8,冷蓉,,,,Y,国家级,中国山水画,125,,240,,,冷蓉/Leng_Rong_14.jpg,5000
+145,柳江小镇,8,冷蓉,,,,Y,国家级,中国山水画,98,,180,,,冷蓉/Leng_Rong_15.jpg,5000
+146,羌山妙语话家园,8,冷蓉,,,,Y,国家级,中国山水画,190,,230,,,冷蓉/Leng_Rong_16.jpg,5000
+147,更喜岷山千里雪,8,冷蓉,,,,Y,国家级,中国山水画,125,,200,,,冷蓉/Leng_Rong_17.jpg,5000
+148,蜀山画意浓,8,冷蓉,,,,Y,国家级,中国山水画,98,,180,,,冷蓉/Leng_Rong_18.jpg,5000
+166,园林印象,10,缪玲,,,,Y,国家级,中国花鸟画,240,,120,,,缪玲/Miu_Ling_01.jpg,8000
+167,金陵寻梦,10,缪玲,,,,Y,国家级,中国花鸟画,180,,96,,,缪玲/Miu_Ling_02.jpg,8000
+168,乡愁,10,缪玲,,,,Y,国家级,中国花鸟画,180,,180,,,缪玲/Miu_Ling_03.jpg,8000
+169,故园清梦,10,缪玲,,,,Y,国家级,中国花鸟画,90,,90,,,缪玲/Miu_Ling_04.jpg,8000
+171,秋 语,10,缪玲,,,,Y,国家级,中国花鸟画,240,,160,,,缪玲/Miu_Ling_06.jpg,8000
+172,秋韵,10,缪玲,,,,Y,国家级,中国花鸟画,195,,180,,,缪玲/Miu_Ling_07.jpg,8000
+173,山间语,10,缪玲,,,,Y,国家级,中国花鸟画,136,,98,,,缪玲/Miu_Ling_08.jpg,8000
+174,藏家小景,10,缪玲,,,,Y,国家级,中国花鸟画,68,,48,,,缪玲/Miu_Ling_09.jpg,8000
+175,园林赋,10,缪玲,,,,Y,国家级,中国花鸟画,180,,97,,,缪玲/Miu_Ling_10.jpg,8000
+176,桃花源记,10,缪玲,,,,Y,国家级,中国花鸟画,136,,98,,,缪玲/Miu_Ling_11.jpg,8000
+177,园林深处待人归,10,缪玲,,,,Y,国家级,中国花鸟画,180,,97,,,缪玲/Miu_Ling_12.jpg,8000
+178,游园惊梦,10,缪玲,,,,Y,国家级,中国花鸟画,180,,97,,,缪玲/Miu_Ling_13.jpg,8000
+179,园中秋景,10,缪玲,,,,Y,国家级,中国花鸟画,65,,57,,,缪玲/Miu_Ling_14.jpg,8000
+180,远上寒石径斜,10,缪玲,,,,Y,国家级,中国花鸟画,136,,68,,,缪玲/Miu_Ling_15.jpg,8000
+181,沁园,10,缪玲,,,,Y,国家级,中国花鸟画,78,,68,,,缪玲/Miu_Ling_16.jpg,8000
+182,林中音,10,缪玲,,,,Y,国家级,中国花鸟画,58,,60,,,缪玲/Miu_Ling_17.jpg,8000
+183,寒园,10,缪玲,,,,Y,国家级,中国花鸟画,68,,136,,,缪玲/Miu_Ling_18.jpg,8000
+184,故园,10,缪玲,,,,Y,国家级,中国花鸟画,68,,136,,,缪玲/Miu_Ling_19.jpg,8000
+185,园中亭,10,缪玲,,,,Y,国家级,中国花鸟画,58,,60,,,缪玲/Miu_Ling_20.jpg,8000
+186,蝶恋花,10,缪玲,,,,Y,国家级,中国花鸟画,68,,68,,,缪玲/Miu_Ling_21.jpg,8000
+187,花鸟团扇,10,缪玲,,,,Y,国家级,中国花鸟画,42,,42,,,缪玲/Miu_Ling_22.jpg,8000
+188,蝶恋花,10,缪玲,,,,Y,国家级,中国花鸟画,42,,42,,,缪玲/Miu_Ling_23.jpg,8000
+236,清赏,13,吴洪涛,,,,Y,国家级,中国花鸟画,54,,70,,,吴洪涛/Wu_Hong_Tao_02.jpg,16329.6
+238,神存富贵,13,吴洪涛,,,,Y,国家级,中国花鸟画,46,,138,,,吴洪涛/Wu_Hong_Tao_04.jpg,27423.36
+239,相逢何处不依依,13,吴洪涛,,,,Y,国家级,中国花鸟画,48,,168,,,吴洪涛/Wu_Hong_Tao_05.jpg,34836.48
+240,清涟,13,吴洪涛,,,,Y,国家级,中国花鸟画,35,,35,,,吴洪涛/Wu_Hong_Tao_06.jpg,5292
+242,轻步临波,13,吴洪涛,,,,Y,国家级,中国花鸟画,48,,48,,,吴洪涛/Wu_Hong_Tao_08.jpg,9953.28
+245,相携何从,13,吴洪涛,,,,Y,国家级,中国花鸟画,141,,48,,,吴洪涛/Wu_Hong_Tao_11.jpg,29237.76
+246,几生修得到梅花,13,吴洪涛,,,,Y,国家级,中国花鸟画,188,,169,,,吴洪涛/Wu_Hong_Tao_12.jpg,137255.04
+247,何须惆怅忆芳容,13,吴洪涛,,,,Y,国家级,中国花鸟画,48,,168,,,吴洪涛/Wu_Hong_Tao_13.jpg,34836.48
+248,君子林,13,吴洪涛,,,,Y,国家级,中国花鸟画,211,,88,,,吴洪涛/Wu_Hong_Tao_14.jpg,80213.76
+249,忽闻疏雨打新荷,13,吴洪涛,,,,Y,国家级,中国花鸟画,48,,68,,,吴洪涛/Wu_Hong_Tao_15.jpg,14100.48
+250,护花千里系长藤,13,吴洪涛,,,,Y,国家级,中国花鸟画,198,,48,,,吴洪涛/Wu_Hong_Tao_16.jpg,41057.28
+1,马兰的辉煌——我国第一颗原子弹爆炸成功,15,筱慧,,,,Y,国家级,水粉画,80,,60,,,筱慧/Xiao_Hui_01.jpg,5000
+2,路旁小景,15,筱慧,,,,Y,国家级,水粉画,53,,37,,,筱慧/Xiao_Hui_02.jpg,5000
+3,松坪沟白蜡下海,15,筱慧,,,,Y,国家级,水粉画,52,,37,,,筱慧/Xiao_Hui_03.jpg,5000
+4,松坪沟小景,15,筱慧,,,,Y,国家级,水粉画,52,,37,,,筱慧/Xiao_Hui_04.jpg,5000
+5,松坪沟白蜡湖,15,筱慧,,,,Y,国家级,水粉画,52,,37,,,筱慧/Xiao_Hui_05.jpg,5000
+6,松坪沟秋韵,15,筱慧,,,,Y,国家级,水粉画,52,,37,,,筱慧/Xiao_Hui_06.jpg,5000
+7,秋韵浓,15,筱慧,,,,Y,国家级,水粉画,52,,37,,,筱慧/Xiao_Hui_07.jpg,5000
+8,松坪沟红桦树,15,筱慧,,,,Y,国家级,水粉画,52,,37,,,筱慧/Xiao_Hui_08.jpg,5000
+9,溪  流,15,筱慧,,,,Y,国家级,水粉画,52,,37,,,筱慧/Xiao_Hui_09.jpg,5000
+10,薰衣草盛开的地方,15,筱慧,,,,Y,国家级,水粉画,52,,37,,,筱慧/Xiao_Hui_10.jpg,5000
+11,光雾山晨曦,15,筱慧,,,,Y,国家级,水粉画,60,,80,,,筱慧/Xiao_Hui_11.jpg,5000
+12,光雾山之春,15,筱慧,,,,Y,国家级,水粉画,30,,40,,,筱慧/Xiao_Hui_12.jpg,5000
+13,光雾山之夏,15,筱慧,,,,Y,国家级,水粉画,30,,40,,,筱慧/Xiao_Hui_13.jpg,5000
+14,光雾山之秋,15,筱慧,,,,Y,国家级,水粉画,30,,40,,,筱慧/Xiao_Hui_14.jpg,5000
+15,光雾山之冬,15,筱慧,,,,Y,国家级,水粉画,30,,40,,,筱慧/Xiao_Hui_15.jpg,5000
+16,九寨沟秋色,15,筱慧,,,,Y,国家级,水粉画,40,,60,,,筱慧/Xiao_Hui_16.jpg,5000
+17,九寨沟珍珠滩,15,筱慧,,,,Y,国家级,水粉画,40,,60,,,筱慧/Xiao_Hui_17.jpg,5000
+18,绿水白石,15,筱慧,,,,Y,国家级,水彩画,100,,160,,,筱慧/Xiao_Hui_18.jpg,5000
+19,聚宝盆,15,筱慧,,,,Y,国家级,油画,60,,80,,,筱慧/Xiao_Hui_19.jpg,5000
+20,装饰花卉,15,筱慧,,,,Y,国家级,水粉画,40,,60,,,筱慧/Xiao_Hui_20.jpg,5000
+21,鱼塘春色,15,筱慧,,,,Y,国家级,油画,60,,80,,,筱慧/Xiao_Hui_21.jpg,5000
+22,花  卉,15,筱慧,,,,Y,国家级,油画,60,,80,,,筱慧/Xiao_Hui_22.jpg,5000
+23,葵花向阳,15,筱慧,,,,Y,国家级,水彩画,52,,37,,,筱慧/Xiao_Hui_23.jpg,5000
+24,成都市花-芙蓉花,15,筱慧,,,,Y,国家级,水彩画,100,,160,,,筱慧/Xiao_Hui_24.jpg,5000
+25,虞美人,15,筱慧,,,,Y,国家级,水彩画,52,,37,,,筱慧/Xiao_Hui_25.jpg,5000
+26,三角梅,15,筱慧,,,,Y,国家级,水粉画,52,,37,,,筱慧/Xiao_Hui_26.jpg,5000
+27,硕果累累,15,筱慧,,,,Y,国家级,水粉画,52,,37,,,筱慧/Xiao_Hui_27.jpg,5000
+28,,16,徐小丹,1982,,,Y,国家级,油画,60,,80,,,徐小丹/Xu_Xiao_Dan_01.jpg,4000
+29,,16,徐小丹,1982,,,Y,国家级,油画,60,,80,,,徐小丹/Xu_Xiao_Dan_02.jpg,4000
+30,,16,徐小丹,1982,,,Y,国家级,油画,60,,80,,,徐小丹/Xu_Xiao_Dan_03.jpg,4000
+31,,16,徐小丹,1982,,,Y,国家级,油画,60,,80,,,徐小丹/Xu_Xiao_Dan_04.jpg,4000
+32,,16,徐小丹,1982,,,Y,国家级,油画,60,,80,,,徐小丹/Xu_Xiao_Dan_05.jpg,4000
+33,,16,徐小丹,1982,,,Y,国家级,油画,60,,80,,,徐小丹/Xu_Xiao_Dan_06.jpg,4000
+34,,16,徐小丹,1982,,,Y,国家级,油画,60,,80,,,徐小丹/Xu_Xiao_Dan_07.jpg,4000
+35,,16,徐小丹,1982,,,Y,国家级,油画,60,,80,,,徐小丹/Xu_Xiao_Dan_08.jpg,4000
+36,,16,徐小丹,1982,,,Y,国家级,油画,60,,80,,,徐小丹/Xu_Xiao_Dan_09.jpg,4000
+37,,16,徐小丹,1982,,,Y,国家级,油画,60,,80,,,徐小丹/Xu_Xiao_Dan_10.jpg,4000
+38,,16,徐小丹,1982,,,Y,国家级,油画,60,,80,,,徐小丹/Xu_Xiao_Dan_11.jpg,4000
+39,,16,徐小丹,1982,,,Y,国家级,油画,60,,80,,,徐小丹/Xu_Xiao_Dan_12.jpg,4000
+40,,16,徐小丹,1982,,,Y,国家级,油画,60,,80,,,徐小丹/Xu_Xiao_Dan_13.jpg,4000
+41,,16,徐小丹,1982,,,Y,国家级,油画,60,,80,,,徐小丹/Xu_Xiao_Dan_14.jpg,4000
+42,,16,徐小丹,1982,,,Y,国家级,油画,60,,80,,,徐小丹/Xu_Xiao_Dan_15.jpg,4000
+43,,16,徐小丹,1982,,,Y,国家级,油画,60,,80,,,徐小丹/Xu_Xiao_Dan_16.jpg,4000
+44,,16,徐小丹,1982,,,Y,国家级,油画,60,,80,,,徐小丹/Xu_Xiao_Dan_17.jpg,4000
+45,,16,徐小丹,1982,,,Y,国家级,油画,60,,80,,,徐小丹/Xu_Xiao_Dan_18.jpg,4000
+46,,16,徐小丹,1982,,,Y,国家级,油画,60,,80,,,徐小丹/Xu_Xiao_Dan_19.jpg,4000
+47,工业物语之一,17,杨绍刚,,,,Y,国家级,布面油画,60,,80,,,杨绍刚/Yang_Shao_Gang_01.jpg,20000
+48,工业物语之二,17,杨绍刚,,,,Y,国家级,布面油画,70,,70,,,杨绍刚/Yang_Shao_Gang_02.jpg,20000
+49,工业物语之三,17,杨绍刚,,,,Y,国家级,布面油画,70,,70,,,杨绍刚/Yang_Shao_Gang_03.jpg,20000
+50,工业物语之四,17,杨绍刚,,,,Y,国家级,布面油画,70,,70,,,杨绍刚/Yang_Shao_Gang_04.jpg,20000
+51,跨  越,17,杨绍刚,,,,Y,国家级,布面油画,190,,110,,,杨绍刚/Yang_Shao_Gang_05.jpg,20000
+52,落  红,17,杨绍刚,,,,Y,国家级,布面油画,130,,110,,,杨绍刚/Yang_Shao_Gang_06.jpg,20000
+53,沐  光,17,杨绍刚,,,,Y,国家级,布面油画,160,,120,,,杨绍刚/Yang_Shao_Gang_07.jpg,20000
+54,弃场系列之厂院,17,杨绍刚,,,,Y,国家级,布面油画,60,,80,,,杨绍刚/Yang_Shao_Gang_08.jpg,20000
+55,弃场系列之晨光,17,杨绍刚,,,,Y,国家级,布面油画,60,,80,,,杨绍刚/Yang_Shao_Gang_09.jpg,20000
+56,太阳雨系列之残阳如血,17,杨绍刚,,,,Y,国家级,布面油画,60,,80,,,杨绍刚/Yang_Shao_Gang_10.jpg,20000
+57,太阳雨系列之晨光,17,杨绍刚,,,,Y,国家级,布面油画,120,,120,,,杨绍刚/Yang_Shao_Gang_11.jpg,20000
+58,太阳雨系列之倒影,17,杨绍刚,,,,Y,国家级,布面油画,60,,80,,,杨绍刚/Yang_Shao_Gang_12.jpg,20000
+59,太阳雨系列之辉煌都市,17,杨绍刚,,,,Y,国家级,布面油画,120,,120,,,杨绍刚/Yang_Shao_Gang_13.jpg,20000
+60,太阳雨系列之落红,17,杨绍刚,,,,Y,国家级,布面油画,60,,80,,,杨绍刚/Yang_Shao_Gang_14.jpg,20000
+61,太阳雨系列之夕阳,17,杨绍刚,,,,Y,国家级,布面油画,60,,80,,,杨绍刚/Yang_Shao_Gang_15.jpg,20000
+62,西藏壁画印象之一,18,臧跃军,1958,,,Y,国家级,中国画,450,,250,,,臧跃军/Zang_Yue_Jun_01.jpg,5000
+63,西藏壁画印象之二,18,臧跃军,1958,,,Y,国家级,中国画,136,,68,,,臧跃军/Zang_Yue_Jun_02.jpg,5000
+64,西藏壁画印象之三,18,臧跃军,1958,,,Y,国家级,中国画,136,,68,,,臧跃军/Zang_Yue_Jun_03.jpg,5000
+65,西藏壁画印象之四,18,臧跃军,1958,,,Y,国家级,中国画,198,,98,,,臧跃军/Zang_Yue_Jun_04.jpg,5000
+66,西藏壁画印象之五,18,臧跃军,1958,,,Y,国家级,中国画,98,,68,,,臧跃军/Zang_Yue_Jun_05.jpg,5000
+67,西藏壁画印象之六,18,臧跃军,1958,,,Y,国家级,中国画,98,,68,,,臧跃军/Zang_Yue_Jun_06.jpg,5000
+68,西藏壁画印象之七,18,臧跃军,1958,,,Y,国家级,中国画,58,,48,,,臧跃军/Zang_Yue_Jun_07.jpg,5000
+69,西藏壁画印象之九,18,臧跃军,1958,,,Y,国家级,中国画,98,,98,,,臧跃军/Zang_Yue_Jun_08.jpg,5000
+70,西藏壁画系列之二,18,臧跃军,1958,,,Y,国家级,中国画,68,,47,,,臧跃军/Zang_Yue_Jun_09.jpg,5000
+71,西藏壁画系列之三,18,臧跃军,1958,,,Y,国家级,中国画,68,,47,,,臧跃军/Zang_Yue_Jun_10.jpg,5000
+72,西藏壁画系列之四,18,臧跃军,1958,,,Y,国家级,中国画,68,,47,,,臧跃军/Zang_Yue_Jun_11.jpg,5000
+73,西藏壁画系列之六,18,臧跃军,1958,,,Y,国家级,中国画,68,,47,,,臧跃军/Zang_Yue_Jun_12.jpg,5000
+74,佛眼中的布达拉,18,臧跃军,1958,,,Y,国家级,中国画,200,,200,,,臧跃军/Zang_Yue_Jun_13.jpg,5000
+75,心灵佛光四一,18,臧跃军,1958,,,Y,国家级,中国画,86.5,,125,,,臧跃军/Zang_Yue_Jun_14.jpg,5000
+76,新疆壁画印象（童子）,18,臧跃军,1958,,,Y,国家级,中国画,138,,68,,,臧跃军/Zang_Yue_Jun_15.jpg,5000
+77,神牛系列之一,18,臧跃军,1958,,,Y,国家级,中国画,140,,76,,,臧跃军/Zang_Yue_Jun_16.jpg,5000
+78,神牛系列之二,18,臧跃军,1958,,,Y,国家级,中国画,76,,68,,,臧跃军/Zang_Yue_Jun_17.jpg,5000
+79,神牛系列之三,18,臧跃军,1958,,,Y,国家级,中国画,140,,76,,,臧跃军/Zang_Yue_Jun_18.jpg,5000
+80,神牛系列之五,18,臧跃军,1958,,,Y,国家级,中国画,140,,76,,,臧跃军/Zang_Yue_Jun_19.jpg,5000
+81,牧  归,18,臧跃军,1958,,,Y,国家级,中国画,46,,69,,,臧跃军/Zang_Yue_Jun_20.jpg,5000
+82,牧  归,18,臧跃军,1958,,,Y,国家级,中国画,68,,47,,,臧跃军/Zang_Yue_Jun_21.jpg,5000
+83,牧归之二,18,臧跃军,1958,,,Y,国家级,中国画,198,,98,,,臧跃军/Zang_Yue_Jun_22.jpg,5000
+84,风情高原,18,臧跃军,1958,,,Y,国家级,中国画,68,,46,,,臧跃军/Zang_Yue_Jun_23.jpg,5000
+85,高原风情,18,臧跃军,1958,,,Y,国家级,中国画,180,,200,,,臧跃军/Zang_Yue_Jun_24.jpg,5000
+86,雄  风,18,臧跃军,1958,,,Y,国家级,中国画,100,,54,,,臧跃军/Zang_Yue_Jun_25.jpg,5000
+87,小画系列之一,18,臧跃军,1958,,,Y,国家级,中国画,23,,80,,,臧跃军/Zang_Yue_Jun_26.jpg,5000
+88,小画系列之二,18,臧跃军,1958,,,Y,国家级,中国画,23,,80,,,臧跃军/Zang_Yue_Jun_27.jpg,5000
+89,放牧图,18,臧跃军,1958,,,Y,国家级,中国画,50,,76,,,臧跃军/Zang_Yue_Jun_28.jpg,5000
+90,风雪牦牛之一,18,臧跃军,1958,,,Y,国家级,中国画,97,,180,,,臧跃军/Zang_Yue_Jun_29.jpg,5000
+91,漫漫风雪路,18,臧跃军,1958,,,Y,国家级,中国画,505,,192,,,臧跃军/Zang_Yue_Jun_30.jpg,5000
+108,梳妆图,20,张学万,1958,,,Y,国家级,没骨画,68,,68,,,张学万/Zhang_Xue_Wan_01.jpg,12484.8
+109,丽人行诗意图,20,张学万,1958,,,Y,国家级,没骨画,68,,126,,,张学万/Zhang_Xue_Wan_02.jpg,23133.6
+110,春风得意丽人行,20,张学万,1958,,,Y,国家级,没骨画,136,,68,,,张学万/Zhang_Xue_Wan_03.jpg,24969.6
+111,杜工部丽人行诗意图,20,张学万,1958,,,Y,国家级,没骨画,68,,136,,,张学万/Zhang_Xue_Wan_04.jpg,24969.6
+112,丽人行诗意图,20,张学万,1958,,,Y,国家级,没骨画,68,,136,,,张学万/Zhang_Xue_Wan_05.jpg,24969.6
+113,薛涛吟句,20,张学万,1958,,,Y,国家级,没骨画,68,,68,,,张学万/Zhang_Xue_Wan_06.jpg,12484.8
+114,丽人行诗意图,20,张学万,1958,,,Y,国家级,没骨画,68,,136,,,张学万/Zhang_Xue_Wan_07.jpg,24969.6
+115,丽人行诗意图,20,张学万,1958,,,Y,国家级,没骨画,68,,136,,,张学万/Zhang_Xue_Wan_08.jpg,24969.6
+116,马球图,20,张学万,1958,,,Y,国家级,没骨画,68,,136,,,张学万/Zhang_Xue_Wan_09.jpg,24969.6
+117,马球图,20,张学万,1958,,,Y,国家级,没骨画,68,,136,,,张学万/Zhang_Xue_Wan_10.jpg,24969.6
+118,游园图,20,张学万,1958,,,Y,国家级,没骨画,68,,68,,,张学万/Zhang_Xue_Wan_11.jpg,12484.8
+119,唐人马球图,20,张学万,1958,,,Y,国家级,没骨画,68,,136,,,张学万/Zhang_Xue_Wan_12.jpg,24969.6
+120,游春图,20,张学万,1958,,,Y,国家级,没骨画,68,,136,,,张学万/Zhang_Xue_Wan_13.jpg,24969.6
+121,阅稿图,20,张学万,1958,,,Y,国家级,没骨画,68,,100,,,张学万/Zhang_Xue_Wan_14.jpg,18360
+122,杜工部丽人行诗意图,20,张学万,1958,,,Y,国家级,没骨画,68,,136,,,张学万/Zhang_Xue_Wan_15.jpg,24969.6
+123,雪中情,20,张学万,1958,,,Y,国家级,没骨画,68,,136,,,张学万/Zhang_Xue_Wan_16.jpg,24969.6
+124,郊游图,20,张学万,1958,,,Y,国家级,没骨画,68,,136,,,张学万/Zhang_Xue_Wan_17.jpg,24969.6
+125,丽人行诗意图,20,张学万,1958,,,Y,国家级,没骨画,98,,198,,,张学万/Zhang_Xue_Wan_18.jpg,52390.8
+126,马球图,20,张学万,1958,,,Y,国家级,没骨画,98,,178,,,张学万/Zhang_Xue_Wan_19.jpg,47098.8
+127,踏春图,20,张学万,1958,,,Y,国家级,没骨画,68,,136,,,张学万/Zhang_Xue_Wan_20.jpg,24969.6
+128,大吉图,20,张学万,1958,,,Y,国家级,没骨画,366,,144,,,张学万/Zhang_Xue_Wan_21.jpg,142300.8
+129,等待春的颜色,21,赵学超,1974,,,Y,国家级,中国画,120,,138,,,赵学超/Zhao_Xue_Chao_01.jpg,59616
+130,格桑花的心思,21,赵学超,1974,,,Y,国家级,中国画,100,,100,,,赵学超/Zhao_Xue_Chao_02.jpg,36000
+131,小憩,21,赵学超,1974,,,Y,国家级,中国画,120,,138,,,赵学超/Zhao_Xue_Chao_03.jpg,59616
+132,朝圣,21,赵学超,1974,,,Y,国家级,中国画,60,,60,,,赵学超/Zhao_Xue_Chao_04.jpg,12960
+133,行,21,赵学超,1974,,,Y,国家级,中国画,80,,100,,,赵学超/Zhao_Xue_Chao_05.jpg,28800
+134,圣山,21,赵学超,1974,,,Y,国家级,中国画,100,,100,,,赵学超/Zhao_Xue_Chao_06.jpg,36000
+135,卓玛的羊,21,赵学超,1974,,,Y,国家级,中国画,60,,80,,,赵学超/Zhao_Xue_Chao_07.jpg,17280
+136,牧场,21,赵学超,1974,,,Y,国家级,中国画,80,,100,,,赵学超/Zhao_Xue_Chao_08.jpg,28800
+137,瞬间,21,赵学超,1974,,,Y,国家级,中国画,80,,100,,,赵学超/Zhao_Xue_Chao_09.jpg,28800
+138,奔跑,21,赵学超,1974,,,Y,国家级,中国画,120,,120,,,赵学超/Zhao_Xue_Chao_10.jpg,51840
+139,奔跑的牦牛,21,赵学超,1974,,,Y,国家级,中国画,50,,60,,,赵学超/Zhao_Xue_Chao_11.jpg,10800
+140,牛气冲天,21,赵学超,1974,,,Y,国家级,中国画,40,,50,,,赵学超/Zhao_Xue_Chao_12.jpg,7200
+141,藏族肖像,21,赵学超,1974,,,Y,国家级,中国画,40,,40,,,赵学超/Zhao_Xue_Chao_13.jpg,5760
+142,肖像,21,赵学超,1974,,,Y,国家级,中国画,40,,40,,,赵学超/Zhao_Xue_Chao_14.jpg,5760
+143,肖像,21,赵学超,1974,,,Y,国家级,中国画,40,,50,,,赵学超/Zhao_Xue_Chao_15.jpg,7200
+144,麻花辫,21,赵学超,1974,,,Y,国家级,中国画,75,,110,,,赵学超/Zhao_Xue_Chao_16.jpg,29700
+145,纸鹤故事之一,21,赵学超,1974,,,Y,国家级,中国画,85,,85,,,赵学超/Zhao_Xue_Chao_17.jpg,26010
+146,放飞的纸鹤,21,赵学超,1974,,,Y,国家级,中国画,100,,100,,,赵学超/Zhao_Xue_Chao_18.jpg,36000
+147,花儿为谁红,21,赵学超,1974,,,Y,国家级,中国画,60,,80,,,赵学超/Zhao_Xue_Chao_19.jpg,17280
+148,窗外,21,赵学超,1974,,,Y,国家级,中国画,80,,100,,,赵学超/Zhao_Xue_Chao_20.jpg,28800
+149,花季,21,赵学超,1974,,,Y,国家级,中国画,80,,100,,,赵学超/Zhao_Xue_Chao_21.jpg,28800
+150,小萌,21,赵学超,1974,,,Y,国家级,中国画,60,,70,,,赵学超/Zhao_Xue_Chao_22.jpg,15120

datasets/test/README.md

@@ -0,0 +1 @@
+用于存放测试图片

datasets/train/README.md

@@ -0,0 +1 @@
+用于存放测试图片

model_data/automl_v2.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a0e54abb42887c0e491c22f81be00032ee74479a11476fe0553c4055d30b8290
+size 310939

model_data/cls_classes.txt

@@ -0,0 +1,2 @@
+cat
+dog

requirements.txt

@@ -0,0 +1,9 @@
+scipy==1.2.1
+numpy==1.17.0
+matplotlib==3.1.2
+tqdm==4.60.0
+Pillow==8.2.0
+h5py==2.10.0
+pandas==1.2.4
+sklearn
+pickle

utils/__init__.py

@@ -0,0 +1 @@
+#

utils/callbacks.py

@@ -0,0 +1,70 @@
+import datetime
+import os
+
+import torch
+import matplotlib
+matplotlib.use('Agg')
+import scipy.signal
+from matplotlib import pyplot as plt
+from torch.utils.tensorboard import SummaryWriter
+
+
+class LossHistory():
+    def __init__(self, log_dir, model, input_shape):
+        time_str        = datetime.datetime.strftime(datetime.datetime.now(),'%Y_%m_%d_%H_%M_%S')
+        self.log_dir    = os.path.join(log_dir, "loss_" + str(time_str))
+        self.losses     = []
+        self.val_loss   = []
+        
+        os.makedirs(self.log_dir)
+        self.writer     = SummaryWriter(self.log_dir)
+        try:
+            dummy_input     = torch.randn(2, 3, input_shape[0], input_shape[1])
+            self.writer.add_graph(model, dummy_input)
+        except:
+            pass
+
+    def append_loss(self, epoch, loss, val_loss):
+        if not os.path.exists(self.log_dir):
+            os.makedirs(self.log_dir)
+
+        self.losses.append(loss)
+        self.val_loss.append(val_loss)
+
+        with open(os.path.join(self.log_dir, "epoch_loss.txt"), 'a') as f:
+            f.write(str(loss))
+            f.write("\n")
+        with open(os.path.join(self.log_dir, "epoch_val_loss.txt"), 'a') as f:
+            f.write(str(val_loss))
+            f.write("\n")
+
+        self.writer.add_scalar('loss', loss, epoch)
+        self.writer.add_scalar('val_loss', val_loss, epoch)
+        self.loss_plot()
+
+    def loss_plot(self):
+        iters = range(len(self.losses))
+
+        plt.figure()
+        plt.plot(iters, self.losses, 'red', linewidth = 2, label='train loss')
+        plt.plot(iters, self.val_loss, 'coral', linewidth = 2, label='val loss')
+        try:
+            if len(self.losses) < 25:
+                num = 5
+            else:
+                num = 15
+            
+            plt.plot(iters, scipy.signal.savgol_filter(self.losses, num, 3), 'green', linestyle = '--', linewidth = 2, label='smooth train loss')
+            plt.plot(iters, scipy.signal.savgol_filter(self.val_loss, num, 3), '#8B4513', linestyle = '--', linewidth = 2, label='smooth val loss')
+        except:
+            pass
+
+        plt.grid(True)
+        plt.xlabel('Epoch')
+        plt.ylabel('Loss')
+        plt.legend(loc="upper right")
+
+        plt.savefig(os.path.join(self.log_dir, "epoch_loss.png"))
+
+        plt.cla()
+        plt.close("all")

utils/dataloader.py

@@ -0,0 +1,193 @@
+from matplotlib import artist
+import cv2
+import numpy as np
+import torch
+import torch.utils.data as data
+import pandas as pd
+import os
+from PIL import Image
+
+from .utils import cvtColor, preprocess_input
+from .utils_aug import CenterCrop, ImageNetPolicy, RandomResizedCrop, Resize
+
+
+class DataGenerator(data.Dataset):
+    def __init__(self, annotation_lines, input_shape, random=True, autoaugment_flag=True):
+        self.artwork_data       = annotation_lines
+        self.input_shape        = input_shape
+        self.random             = random
+        #------------------------------#
+        #   是否使用数据增强
+        #------------------------------#
+        self.autoaugment_flag   = autoaugment_flag
+        if self.autoaugment_flag:
+            self.resize_crop = RandomResizedCrop(input_shape)
+            self.policy      = ImageNetPolicy()
+            
+            self.resize      = Resize(input_shape[0] if input_shape[0] == input_shape[1] else input_shape)
+            self.center_crop = CenterCrop(input_shape)
+        self.all_features    = self.get_all_features(self.artwork_data)
+
+    def __len__(self):
+        return self.artwork_data.shape[0]
+
+    def __getitem__(self, index):
+        # 从数据集中获取图像地址
+        annotation_path = self.artwork_data['Duration (s)'][index]
+        annotation_path = os.path.join('datasets/archive/Dataset', annotation_path)
+        image = Image.open(annotation_path)
+        #------------------------------#
+        #   读取图像并转换成RGB图像
+        #------------------------------#
+        image   = cvtColor(image)
+        if self.autoaugment_flag:
+            image = self.AutoAugment(image, random=self.random)
+        else:
+            image = self.get_random_data(image, self.input_shape, random=self.random)
+        # 去除价格特征
+        other_features = self.all_features.drop(labels='Prices', axis=1)
+        # 取其它特征作为输入
+        other_features = other_features.iloc[index].values
+        other_features = np.resize(np.array(other_features, dtype=np.float32), self.input_shape)
+        other_features = np.expand_dims(other_features, 0)
+        image          = np.transpose(preprocess_input(np.array(image, dtype=np.float32)), [2, 0, 1])
+        all_features   = np.concatenate((image, other_features), axis=0)
+        # 从数据集中获取价格标签
+        y = np.expand_dims(np.array(self.all_features['Prices'][index]), axis=-1)
+        return all_features, y
+
+    def rand(self, a=0, b=1):
+        return np.random.rand()*(b-a) + a
+
+    '''
+    @description: 
+    @param {*} self
+    @param {*} all_features
+    @return {*} 数据集预处理
+    '''        
+    def get_all_features(self, all_features):
+        all_features = all_features.iloc[:, [2,4,5,7,8,9,10,12,16]]
+        all_features = all_features.copy()
+        numeric_features = all_features.dtypes[all_features.dtypes != 'object'].index
+        all_features[numeric_features] = all_features[numeric_features].apply(
+            lambda x: (x - x.mean()) / (x.std()))
+        # 标准化后，每个特征的均值变为0，所以可以直接用0来替换缺失值
+        all_features[numeric_features] = all_features[numeric_features].fillna(0)
+        # dummy_na=True将缺失值也当作合法的特征值并为其创建指示特征
+        all_features = pd.get_dummies(all_features, dummy_na=True)
+        return all_features
+
+    def get_random_data(self, image, input_shape, jitter=.3, hue=.1, sat=1.5, val=1.5, random=True):
+        #------------------------------#
+        #   获得图像的高宽与目标高宽
+        #------------------------------#
+        iw, ih  = image.size
+        h, w    = input_shape
+
+        if not random:
+            scale = min(w/iw, h/ih)
+            nw = int(iw*scale)
+            nh = int(ih*scale)
+            dx = (w-nw)//2
+            dy = (h-nh)//2
+
+            #---------------------------------#
+            #   将图像多余的部分加上灰条
+            #---------------------------------#
+            image       = image.resize((nw,nh), Image.BICUBIC)
+            new_image   = Image.new('RGB', (w,h), (128,128,128))
+            new_image.paste(image, (dx, dy))
+            image_data  = np.array(new_image, np.float32)
+
+            return image_data
+
+        #------------------------------------------#
+        #   对图像进行缩放并且进行长和宽的扭曲
+        #------------------------------------------#
+        new_ar = iw/ih * self.rand(1-jitter,1+jitter) / self.rand(1-jitter,1+jitter)
+        scale = self.rand(.75, 1.5)
+        if new_ar < 1:
+            nh = int(scale*h)
+            nw = int(nh*new_ar)
+        else:
+            nw = int(scale*w)
+            nh = int(nw/new_ar)
+        image = image.resize((nw,nh), Image.BICUBIC)
+
+        #------------------------------------------#
+        #   将图像多余的部分加上灰条
+        #------------------------------------------#
+        dx = int(self.rand(0, w-nw))
+        dy = int(self.rand(0, h-nh))
+        new_image = Image.new('RGB', (w,h), (128,128,128))
+        new_image.paste(image, (dx, dy))
+        image = new_image
+
+        #------------------------------------------#
+        #   翻转图像
+        #------------------------------------------#
+        flip = self.rand()<.5
+        if flip: image = image.transpose(Image.FLIP_LEFT_RIGHT)
+        
+        rotate = self.rand()<.5
+        if rotate: 
+            angle = np.random.randint(-15,15)
+            a,b = w/2,h/2
+            M = cv2.getRotationMatrix2D((a,b),angle,1)
+            image = cv2.warpAffine(np.array(image), M, (w,h), borderValue=[128, 128, 128]) 
+
+        image_data      = np.array(image, np.uint8)
+        #---------------------------------#
+        #   对图像进行色域变换
+        #   计算色域变换的参数
+        #---------------------------------#
+        r               = np.random.uniform(-1, 1, 3) * [hue, sat, val] + 1
+        #---------------------------------#
+        #   将图像转到HSV上
+        #---------------------------------#
+        hue, sat, val   = cv2.split(cv2.cvtColor(image_data, cv2.COLOR_RGB2HSV))
+        dtype           = image_data.dtype
+        #---------------------------------#
+        #   应用变换
+        #---------------------------------#
+        x       = np.arange(0, 256, dtype=r.dtype)
+        lut_hue = ((x * r[0]) % 180).astype(dtype)
+        lut_sat = np.clip(x * r[1], 0, 255).astype(dtype)
+        lut_val = np.clip(x * r[2], 0, 255).astype(dtype)
+
+        image_data = cv2.merge((cv2.LUT(hue, lut_hue), cv2.LUT(sat, lut_sat), cv2.LUT(val, lut_val)))
+        image_data = cv2.cvtColor(image_data, cv2.COLOR_HSV2RGB)
+        return image_data
+    
+    def AutoAugment(self, image, random=True):
+        if not random:
+            image = self.resize(image)
+            image = self.center_crop(image)
+            return image
+
+        #------------------------------------------#
+        #   resize并且随即裁剪
+        #------------------------------------------#
+        image = self.resize_crop(image)
+        
+        #------------------------------------------#
+        #   翻转图像
+        #------------------------------------------#
+        flip = self.rand()<.5
+        if flip: image = image.transpose(Image.FLIP_LEFT_RIGHT)
+        
+        #------------------------------------------#
+        #   随机增强
+        #------------------------------------------#
+        image = self.policy(image)
+        return image
+            
+def detection_collate(batch):
+    images = []
+    targets = []
+    for image, y in batch:
+        images.append(image)
+        targets.append(y)
+    images  = torch.from_numpy(np.array(images)).type(torch.FloatTensor)
+    targets = torch.from_numpy(np.array(targets)).type(torch.FloatTensor)
+    return images, targets

utils/utils.py

@@ -0,0 +1,149 @@
+import math
+from functools import partial
+
+import numpy as np
+import torch
+from PIL import Image
+
+from .utils_aug import resize, center_crop
+
+
+#---------------------------------------------------------#
+#   将图像转换成RGB图像，防止灰度图在预测时报错。
+#   代码仅仅支持RGB图像的预测，所有其它类型的图像都会转化成RGB
+#---------------------------------------------------------#
+def cvtColor(image):
+    if len(np.shape(image)) == 3 and np.shape(image)[2] == 3:
+        return image 
+    else:
+        image = image.convert('RGB')
+        return image 
+
+#---------------------------------------------------#
+#   对输入图像进行resize
+#---------------------------------------------------#
+def letterbox_image(image, size, letterbox_image):
+    w, h = size
+    iw, ih = image.size
+    if letterbox_image:
+        '''resize image with unchanged aspect ratio using padding'''
+        scale = min(w/iw, h/ih)
+        nw = int(iw*scale)
+        nh = int(ih*scale)
+
+        image = image.resize((nw,nh), Image.BICUBIC)
+        new_image = Image.new('RGB', size, (128,128,128))
+        new_image.paste(image, ((w-nw)//2, (h-nh)//2))
+    else:
+        if h == w:
+            new_image = resize(image, h)
+        else:
+            new_image = resize(image, [h ,w])
+        new_image = center_crop(new_image, [h ,w])
+    return new_image
+
+#---------------------------------------------------#
+#   获得类
+#---------------------------------------------------#
+def get_classes(classes_path):
+    with open(classes_path, encoding='utf-8') as f:
+        class_names = f.readlines()
+    class_names = [c.strip() for c in class_names]
+    return class_names, len(class_names)
+
+#----------------------------------------#
+#   预处理训练图片
+#----------------------------------------#
+def preprocess_input(x):
+    x /= 127.5
+    x -= 1.
+    return x
+
+def show_config(**kwargs):
+    print('Configurations:')
+    print('-' * 70)
+    print('|%25s | %40s|' % ('keys', 'values'))
+    print('-' * 70)
+    for key, value in kwargs.items():
+        print('|%25s | %40s|' % (str(key), str(value)))
+    print('-' * 70)
+
+#---------------------------------------------------#
+#   获得学习率
+#---------------------------------------------------#
+def get_lr(optimizer):
+    for param_group in optimizer.param_groups:
+        return param_group['lr']
+
+def weights_init(net, init_type='normal', init_gain=0.02):
+    def init_func(m):
+        classname = m.__class__.__name__
+        if hasattr(m, 'weight') and classname.find('Conv') != -1:
+            if init_type == 'normal':
+                torch.nn.init.normal_(m.weight.data, 0.0, init_gain)
+            elif init_type == 'xavier':
+                torch.nn.init.xavier_normal_(m.weight.data, gain=init_gain)
+            elif init_type == 'kaiming':
+                torch.nn.init.kaiming_normal_(m.weight.data, a=0, mode='fan_in')
+            elif init_type == 'orthogonal':
+                torch.nn.init.orthogonal_(m.weight.data, gain=init_gain)
+            else:
+                raise NotImplementedError('initialization method [%s] is not implemented' % init_type)
+        elif classname.find('BatchNorm2d') != -1:
+            torch.nn.init.normal_(m.weight.data, 1.0, 0.02)
+            torch.nn.init.constant_(m.bias.data, 0.0)
+    print('initialize network with %s type' % init_type)
+    net.apply(init_func)
+
+def get_lr_scheduler(lr_decay_type, lr, min_lr, total_iters, warmup_iters_ratio = 0.05, warmup_lr_ratio = 0.1, no_aug_iter_ratio = 0.05, step_num = 10):
+    def yolox_warm_cos_lr(lr, min_lr, total_iters, warmup_total_iters, warmup_lr_start, no_aug_iter, iters):
+        if iters <= warmup_total_iters:
+            # lr = (lr - warmup_lr_start) * iters / float(warmup_total_iters) + warmup_lr_start
+            lr = (lr - warmup_lr_start) * pow(iters / float(warmup_total_iters), 2) + warmup_lr_start
+        elif iters >= total_iters - no_aug_iter:
+            lr = min_lr
+        else:
+            lr = min_lr + 0.5 * (lr - min_lr) * (
+                1.0 + math.cos(math.pi* (iters - warmup_total_iters) / (total_iters - warmup_total_iters - no_aug_iter))
+            )
+        return lr
+
+    def step_lr(lr, decay_rate, step_size, iters):
+        if step_size < 1:
+            raise ValueError("step_size must above 1.")
+        n       = iters // step_size
+        out_lr  = lr * decay_rate ** n
+        return out_lr
+
+    if lr_decay_type == "cos":
+        warmup_total_iters  = min(max(warmup_iters_ratio * total_iters, 1), 3)
+        warmup_lr_start     = max(warmup_lr_ratio * lr, 1e-6)
+        no_aug_iter         = min(max(no_aug_iter_ratio * total_iters, 1), 15)
+        func = partial(yolox_warm_cos_lr ,lr, min_lr, total_iters, warmup_total_iters, warmup_lr_start, no_aug_iter)
+    else:
+        decay_rate  = (min_lr / lr) ** (1 / (step_num - 1))
+        step_size   = total_iters / step_num
+        func = partial(step_lr, lr, decay_rate, step_size)
+
+    return func
+
+def set_optimizer_lr(optimizer, lr_scheduler_func, epoch):
+    lr = lr_scheduler_func(epoch)
+    for param_group in optimizer.param_groups:
+        param_group['lr'] = lr
+
+def download_weights(backbone, model_dir="./model_data"):
+    import os
+    from torch.hub import load_state_dict_from_url
+    
+    download_urls = {
+        'vgg16'         : 'https://download.pytorch.org/models/vgg16-397923af.pth',
+        'mobilenet'     : 'https://download.pytorch.org/models/mobilenet_v2-b0353104.pth',
+        'resnet50'      : 'https://download.pytorch.org/models/resnet50-19c8e357.pth',
+        'vit'           : 'https://github.com/bubbliiiing/classification-pytorch/releases/download/v1.0/vit-patch_16.pth'
+    }
+    url = download_urls[backbone]
+    
+    if not os.path.exists(model_dir):
+        os.makedirs(model_dir)
+    load_state_dict_from_url(url, model_dir)

utils/utils_aug.py

@@ -0,0 +1,404 @@
+import numbers
+import random
+import math
+import warnings
+import numpy as np
+from PIL import Image, ImageEnhance, ImageOps
+
+
+class ShearX(object):
+    def __init__(self, fillcolor=(128, 128, 128)):
+        self.fillcolor = fillcolor
+
+    def __call__(self, x, magnitude):
+        return x.transform(
+            x.size, Image.AFFINE, (1, magnitude * random.choice([-1, 1]), 0, 0, 1, 0),
+            Image.BICUBIC, fillcolor=self.fillcolor)
+
+
+class ShearY(object):
+    def __init__(self, fillcolor=(128, 128, 128)):
+        self.fillcolor = fillcolor
+
+    def __call__(self, x, magnitude):
+        return x.transform(
+            x.size, Image.AFFINE, (1, 0, 0, magnitude * random.choice([-1, 1]), 1, 0),
+            Image.BICUBIC, fillcolor=self.fillcolor)
+
+
+class TranslateX(object):
+    def __init__(self, fillcolor=(128, 128, 128)):
+        self.fillcolor = fillcolor
+
+    def __call__(self, x, magnitude):
+        return x.transform(
+            x.size, Image.AFFINE, (1, 0, magnitude * x.size[0] * random.choice([-1, 1]), 0, 1, 0),
+            fillcolor=self.fillcolor)
+
+
+class TranslateY(object):
+    def __init__(self, fillcolor=(128, 128, 128)):
+        self.fillcolor = fillcolor
+
+    def __call__(self, x, magnitude):
+        return x.transform(
+            x.size, Image.AFFINE, (1, 0, 0, 0, 1, magnitude * x.size[1] * random.choice([-1, 1])),
+            fillcolor=self.fillcolor)
+
+
+class Rotate(object):
+    # from https://stackoverflow.com/questions/
+    # 5252170/specify-image-filling-color-when-rotating-in-python-with-pil-and-setting-expand
+    def __call__(self, x, magnitude):
+        rot = x.convert("RGBA").rotate(magnitude * random.choice([-1, 1]))
+        return Image.composite(rot, Image.new("RGBA", rot.size, (128,) * 4), rot).convert(x.mode)
+
+
+class Color(object):
+    def __call__(self, x, magnitude):
+        return ImageEnhance.Color(x).enhance(1 + magnitude * random.choice([-1, 1]))
+
+
+class Posterize(object):
+    def __call__(self, x, magnitude):
+        return ImageOps.posterize(x, magnitude)
+
+
+class Solarize(object):
+    def __call__(self, x, magnitude):
+        return ImageOps.solarize(x, magnitude)
+
+
+class Contrast(object):
+    def __call__(self, x, magnitude):
+        return ImageEnhance.Contrast(x).enhance(1 + magnitude * random.choice([-1, 1]))
+
+
+class Sharpness(object):
+    def __call__(self, x, magnitude):
+        return ImageEnhance.Sharpness(x).enhance(1 + magnitude * random.choice([-1, 1]))
+
+
+class Brightness(object):
+    def __call__(self, x, magnitude):
+        return ImageEnhance.Brightness(x).enhance(1 + magnitude * random.choice([-1, 1]))
+
+
+class AutoContrast(object):
+    def __call__(self, x, magnitude):
+        return ImageOps.autocontrast(x)
+
+
+class Equalize(object):
+    def __call__(self, x, magnitude):
+        return ImageOps.equalize(x)
+
+
+class Invert(object):
+    def __call__(self, x, magnitude):
+        return ImageOps.invert(x)
+
+
+class ImageNetPolicy(object):
+    """ Randomly choose one of the best 24 Sub-policies on ImageNet.
+        Example:
+        >>> policy = ImageNetPolicy()
+        >>> transformed = policy(image)
+        Example as a PyTorch Transform:
+        >>> transform = transforms.Compose([
+        >>>     transforms.Resize(256),
+        >>>     ImageNetPolicy(),
+        >>>     transforms.ToTensor()])
+    """
+    def __init__(self, fillcolor=(128, 128, 128)):
+        self.policies = [
+            SubPolicy(0.4, "posterize", 8, 0.6, "rotate", 9, fillcolor),
+            SubPolicy(0.6, "solarize", 5, 0.6, "autocontrast", 5, fillcolor),
+            SubPolicy(0.8, "equalize", 8, 0.6, "equalize", 3, fillcolor),
+            SubPolicy(0.6, "posterize", 7, 0.6, "posterize", 6, fillcolor),
+            SubPolicy(0.4, "equalize", 7, 0.2, "solarize", 4, fillcolor),
+
+            SubPolicy(0.4, "equalize", 4, 0.8, "rotate", 8, fillcolor),
+            SubPolicy(0.6, "solarize", 3, 0.6, "equalize", 7, fillcolor),
+            SubPolicy(0.8, "posterize", 5, 1.0, "equalize", 2, fillcolor),
+            SubPolicy(0.2, "rotate", 3, 0.6, "solarize", 8, fillcolor),
+            SubPolicy(0.6, "equalize", 8, 0.4, "posterize", 6, fillcolor),
+
+            SubPolicy(0.8, "rotate", 8, 0.4, "color", 0, fillcolor),
+            SubPolicy(0.4, "rotate", 9, 0.6, "equalize", 2, fillcolor),
+            SubPolicy(0.0, "equalize", 7, 0.8, "equalize", 8, fillcolor),
+            SubPolicy(0.6, "invert", 4, 1.0, "equalize", 8, fillcolor),
+            SubPolicy(0.6, "color", 4, 1.0, "contrast", 8, fillcolor),
+
+            SubPolicy(0.8, "rotate", 8, 1.0, "color", 2, fillcolor),
+            SubPolicy(0.8, "color", 8, 0.8, "solarize", 7, fillcolor),
+            SubPolicy(0.4, "sharpness", 7, 0.6, "invert", 8, fillcolor),
+            SubPolicy(0.6, "shearX", 5, 1.0, "equalize", 9, fillcolor),
+            SubPolicy(0.4, "color", 0, 0.6, "equalize", 3, fillcolor),
+
+            SubPolicy(0.4, "equalize", 7, 0.2, "solarize", 4, fillcolor),
+            SubPolicy(0.6, "solarize", 5, 0.6, "autocontrast", 5, fillcolor),
+            SubPolicy(0.6, "invert", 4, 1.0, "equalize", 8, fillcolor),
+            SubPolicy(0.6, "color", 4, 1.0, "contrast", 8, fillcolor),
+            SubPolicy(0.8, "equalize", 8, 0.6, "equalize", 3, fillcolor)
+        ]
+
+    def __call__(self, img):
+        policy_idx = random.randint(0, len(self.policies) - 1)
+        return self.policies[policy_idx](img)
+
+    def __repr__(self):
+        return "AutoAugment ImageNet Policy"
+
+class SubPolicy(object):
+    def __init__(self, p1, operation1, magnitude_idx1, p2, operation2, magnitude_idx2, fillcolor=(128, 128, 128)):
+        ranges = {
+            "shearX": np.linspace(0, 0.3, 10),
+            "shearY": np.linspace(0, 0.3, 10),
+            "translateX": np.linspace(0, 150 / 331, 10),
+            "translateY": np.linspace(0, 150 / 331, 10),
+            "rotate": np.linspace(0, 30, 10),
+            "color": np.linspace(0.0, 0.9, 10),
+            "posterize": np.round(np.linspace(8, 4, 10), 0).astype(np.int),
+            "solarize": np.linspace(256, 0, 10),
+            "contrast": np.linspace(0.0, 0.9, 10),
+            "sharpness": np.linspace(0.0, 0.9, 10),
+            "brightness": np.linspace(0.0, 0.9, 10),
+            "autocontrast": [0] * 10,
+            "equalize": [0] * 10,
+            "invert": [0] * 10
+        }
+
+        func = {
+            "shearX": ShearX(fillcolor=fillcolor),
+            "shearY": ShearY(fillcolor=fillcolor),
+            "translateX": TranslateX(fillcolor=fillcolor),
+            "translateY": TranslateY(fillcolor=fillcolor),
+            "rotate": Rotate(),
+            "color": Color(),
+            "posterize": Posterize(),
+            "solarize": Solarize(),
+            "contrast": Contrast(),
+            "sharpness": Sharpness(),
+            "brightness": Brightness(),
+            "autocontrast": AutoContrast(),
+            "equalize": Equalize(),
+            "invert": Invert()
+        }
+
+        self.p1 = p1
+        self.operation1 = func[operation1]
+        self.magnitude1 = ranges[operation1][magnitude_idx1]
+        self.p2 = p2
+        self.operation2 = func[operation2]
+        self.magnitude2 = ranges[operation2][magnitude_idx2]
+
+    def __call__(self, img):
+        if random.random() < self.p1:
+            img = self.operation1(img, self.magnitude1)
+        if random.random() < self.p2:
+            img = self.operation2(img, self.magnitude2)
+        return img
+
+def crop(img, i, j, h, w):
+    """Crop the given PIL Image.
+
+    Args:
+        img (PIL Image): Image to be cropped.
+        i (int): i in (i,j) i.e coordinates of the upper left corner.
+        j (int): j in (i,j) i.e coordinates of the upper left corner.
+        h (int): Height of the cropped image.
+        w (int): Width of the cropped image.
+
+    Returns:
+        PIL Image: Cropped image.
+    """
+    return img.crop((j, i, j + w, i + h))
+
+def resize(img, size, interpolation=Image.BILINEAR):
+    r"""Resize the input PIL Image to the given size.
+
+    Args:
+        img (PIL Image): Image to be resized.
+        size (sequence or int): Desired output size. If size is a sequence like
+            (h, w), the output size will be matched to this. If size is an int,
+            the smaller edge of the image will be matched to this number maintaing
+            the aspect ratio. i.e, if height > width, then image will be rescaled to
+            :math:`\left(\text{size} \times \frac{\text{height}}{\text{width}}, \text{size}\right)`
+        interpolation (int, optional): Desired interpolation. Default is
+            ``PIL.Image.BILINEAR``
+
+    Returns:
+        PIL Image: Resized image.
+    """
+    if isinstance(size, int):
+        w, h = img.size
+        if (w <= h and w == size) or (h <= w and h == size):
+            return img
+        if w < h:
+            ow = size
+            oh = int(size * h / w)
+            return img.resize((ow, oh), interpolation)
+        else:
+            oh = size
+            ow = int(size * w / h)
+            return img.resize((ow, oh), interpolation)
+    else:
+        return img.resize(size[::-1], interpolation)
+
+def center_crop(img, output_size):
+    if isinstance(output_size, numbers.Number):
+        output_size = (int(output_size), int(output_size))
+    w, h = img.size
+    th, tw = output_size
+    i = int(round((h - th) / 2.))
+    j = int(round((w - tw) / 2.))
+    return crop(img, i, j, th, tw)
+
+def resized_crop(img, i, j, h, w, size, interpolation=Image.BILINEAR):
+    """Crop the given PIL Image and resize it to desired size.
+
+    Notably used in :class:`~torchvision.transforms.RandomResizedCrop`.
+
+    Args:
+        img (PIL Image): Image to be cropped.
+        i (int): i in (i,j) i.e coordinates of the upper left corner
+        j (int): j in (i,j) i.e coordinates of the upper left corner
+        h (int): Height of the cropped image.
+        w (int): Width of the cropped image.
+        size (sequence or int): Desired output size. Same semantics as ``resize``.
+        interpolation (int, optional): Desired interpolation. Default is
+            ``PIL.Image.BILINEAR``.
+    Returns:
+        PIL Image: Cropped image.
+    """
+    img = crop(img, i, j, h, w)
+    img = resize(img, size, interpolation)
+    return img
+
+class Resize(object):
+    """Resize the input PIL Image to the given size.
+
+    Args:
+        size (sequence or int): Desired output size. If size is a sequence like
+            (h, w), output size will be matched to this. If size is an int,
+            smaller edge of the image will be matched to this number.
+            i.e, if height > width, then image will be rescaled to
+            (size * height / width, size)
+        interpolation (int, optional): Desired interpolation. Default is
+            ``PIL.Image.BILINEAR``
+    """
+
+    def __init__(self, size, interpolation=Image.BILINEAR):
+        self.size           = size
+        self.interpolation  = interpolation
+
+    def __call__(self, img):
+        """
+        Args:
+            img (PIL Image): Image to be scaled.
+
+        Returns:
+            PIL Image: Rescaled image.
+        """
+        return resize(img, self.size, self.interpolation)
+    
+class CenterCrop(object):
+    """Crops the given PIL Image at the center.
+
+    Args:
+        size (sequence or int): Desired output size of the crop. If size is an
+            int instead of sequence like (h, w), a square crop (size, size) is
+            made.
+    """
+
+    def __init__(self, size):
+        self.size = size
+
+    def __call__(self, img):
+        """
+        Args:
+            img (PIL Image): Image to be cropped.
+
+        Returns:
+            PIL Image: Cropped image.
+        """
+        return center_crop(img, self.size)
+
+class RandomResizedCrop(object):
+    """Crop the given PIL Image to random size and aspect ratio.
+
+    A crop of random size (default: of 0.08 to 1.0) of the original size and a random
+    aspect ratio (default: of 3/4 to 4/3) of the original aspect ratio is made. This crop
+    is finally resized to given size.
+    This is popularly used to train the Inception networks.
+
+    Args:
+        size: expected output size of each edge
+        scale: range of size of the origin size cropped
+        ratio: range of aspect ratio of the origin aspect ratio cropped
+        interpolation: Default: PIL.Image.BILINEAR
+    """
+
+    def __init__(self, size, scale=(0.08, 1.0), ratio=(3. / 4., 4. / 3.), interpolation=Image.BILINEAR):
+        self.size = size
+        if (scale[0] > scale[1]) or (ratio[0] > ratio[1]):
+            warnings.warn("range should be of kind (min, max)")
+
+        self.interpolation = interpolation
+        self.scale = scale
+        self.ratio = ratio
+
+    @staticmethod
+    def get_params(img, scale, ratio):
+        """Get parameters for ``crop`` for a random sized crop.
+
+        Args:
+            img (PIL Image): Image to be cropped.
+            scale (tuple): range of size of the origin size cropped
+            ratio (tuple): range of aspect ratio of the origin aspect ratio cropped
+
+        Returns:
+            tuple: params (i, j, h, w) to be passed to ``crop`` for a random
+                sized crop.
+        """
+        area = img.size[0] * img.size[1]
+
+        for attempt in range(10):
+            target_area = random.uniform(*scale) * area
+            log_ratio = (math.log(ratio[0]), math.log(ratio[1]))
+            aspect_ratio = math.exp(random.uniform(*log_ratio))
+
+            w = int(round(math.sqrt(target_area * aspect_ratio)))
+            h = int(round(math.sqrt(target_area / aspect_ratio)))
+
+            if w <= img.size[0] and h <= img.size[1]:
+                i = random.randint(0, img.size[1] - h)
+                j = random.randint(0, img.size[0] - w)
+                return i, j, h, w
+
+        # Fallback to central crop
+        in_ratio = img.size[0] / img.size[1]
+        if (in_ratio < min(ratio)):
+            w = img.size[0]
+            h = int(round(w / min(ratio)))
+        elif (in_ratio > max(ratio)):
+            h = img.size[1]
+            w = int(round(h * max(ratio)))
+        else:  # whole image
+            w = img.size[0]
+            h = img.size[1]
+        i = (img.size[1] - h) // 2
+        j = (img.size[0] - w) // 2
+        return i, j, h, w
+
+    def __call__(self, img):
+        """
+        Args:
+            img (PIL Image): Image to be cropped and resized.
+
+        Returns:
+            PIL Image: Randomly cropped and resized image.
+        """
+        i, j, h, w = self.get_params(img, self.scale, self.ratio)
+        return resized_crop(img, i, j, h, w, self.size, self.interpolation)

utils/utils_fit.py

@@ -0,0 +1,131 @@
+import os
+from threading import local
+
+import torch
+import torch.nn.functional as F
+from torch import nn
+from tqdm import tqdm
+
+from .utils import get_lr
+
+def log_rmse(outputs, labels, loss):
+    with torch.no_grad():
+        # 将小于1的值设成1，使得取对数时数值更稳定
+        clipped_preds = torch.max(outputs, torch.tensor(1.0))
+        rmse = torch.sqrt(2 * loss(clipped_preds.log(), labels.log()).mean())
+    return rmse
+
+def fit_one_epoch(model_train, model, loss_history, optimizer, epoch, epoch_step, epoch_step_val, gen, gen_val, Epoch, cuda, fp16, scaler, save_period, save_dir, local_rank=0):
+    total_loss      = 0
+    total_rmse      = 0
+
+    val_loss        = 0
+    val_rmse        = 0
+    # 定义损失函数
+    loss            = nn.MSELoss()
+    if local_rank == 0:
+        print('Start Train')
+        pbar = tqdm(total=epoch_step,desc=f'Epoch {epoch + 1}/{Epoch}',postfix=dict,mininterval=0.3)
+    model_train.train()
+    for iteration, batch in enumerate(gen):
+        if iteration >= epoch_step: 
+            break
+        images, targets = batch
+        with torch.no_grad():
+            if cuda:
+                images  = images.cuda(local_rank)
+                targets = targets.cuda(local_rank)
+                
+        #----------------------#
+        #   清零梯度
+        #----------------------#
+        optimizer.zero_grad()
+        if not fp16:
+            #----------------------#
+            #   前向传播
+            #----------------------#
+            outputs     = model_train(images)
+            #----------------------#
+            #   计算损失
+            #----------------------#
+            loss_value  = loss(outputs, targets)
+            loss_value.backward()
+            optimizer.step()
+        else:
+            from torch.cuda.amp import autocast
+            with autocast():
+                #----------------------#
+                #   前向传播
+                #----------------------#
+                outputs     = model_train(images)
+                #----------------------#
+                #   计算损失
+                #----------------------#
+                loss_value  = loss(outputs, targets)
+            #----------------------#
+            #   反向传播
+            #----------------------#
+            scaler.scale(loss_value).backward()
+            scaler.step(optimizer)
+            scaler.update()
+
+        total_loss += loss_value.item()
+        # 计算对数均方根误差
+        with torch.no_grad():
+            rmse = log_rmse(outputs, targets, loss)
+            total_rmse += rmse.item()
+
+        if local_rank == 0:
+            pbar.set_postfix(**{'total_loss': total_loss / (iteration + 1), 
+                                'total_rmse': total_rmse / (iteration + 1), 
+                                'lr'        : get_lr(optimizer)})
+            pbar.update(1)
+
+    if local_rank == 0:
+        pbar.close()
+        print('Finish Train')
+        print('Start Validation')
+        pbar = tqdm(total=epoch_step_val, desc=f'Epoch {epoch + 1}/{Epoch}',postfix=dict,mininterval=0.3)
+    model_train.eval()
+    for iteration, batch in enumerate(gen_val):
+        if iteration >= epoch_step_val:
+            break
+        images, targets = batch
+        with torch.no_grad():
+            if cuda:
+                images  = images.cuda(local_rank)
+                targets = targets.cuda(local_rank)
+
+            optimizer.zero_grad()
+
+            outputs     = model_train(images)
+            loss_value  = loss(outputs, targets)
+            
+            val_loss    += loss_value.item()
+            rmse        = log_rmse(outputs, targets, loss)
+            val_rmse    += rmse.item()
+            
+        if local_rank == 0:
+            pbar.set_postfix(**{'total_loss': val_loss / (iteration + 1),
+                                'total_rmse': val_rmse / (iteration + 1), 
+                                'lr'        : get_lr(optimizer)})
+            pbar.update(1)
+                
+    if local_rank == 0:
+        pbar.close()
+        print('Finish Validation')
+        loss_history.append_loss(epoch + 1, total_loss / epoch_step, val_loss / epoch_step_val)
+        print('Epoch:' + str(epoch + 1) + '/' + str(Epoch))
+        print('Total Loss: %.3f || Val Loss: %.3f ' % (total_loss / epoch_step, val_loss / epoch_step_val))
+        
+        #-----------------------------------------------#
+        #   保存权值
+        #-----------------------------------------------#
+        if (epoch + 1) % save_period == 0 or epoch + 1 == Epoch:
+            torch.save(model.state_dict(), os.path.join(save_dir, "ep%03d-loss%.3f-val_loss%.3f.pth" % (epoch + 1, total_loss / epoch_step, val_loss / epoch_step_val)))
+
+        if len(loss_history.val_loss) <= 1 or (val_loss / epoch_step_val) <= min(loss_history.val_loss):
+            print('Save best model to best_epoch_weights.pth')
+            torch.save(model.state_dict(), os.path.join(save_dir, "best_epoch_weights.pth"))
+            
+        torch.save(model.state_dict(), os.path.join(save_dir, "last_epoch_weights.pth"))

utils/utils_metrics.py

@@ -0,0 +1,120 @@
+import csv
+import os
+
+import matplotlib.pyplot as plt
+import numpy as np
+from PIL import Image
+
+def evaluteTop1_5(classfication, lines, metrics_out_path):
+    correct_1 = 0
+    correct_5 = 0
+    preds   = []
+    labels  = []
+    total = len(lines)
+    for index, line in enumerate(lines):
+        annotation_path = line.split(';')[1].split()[0]
+        x = Image.open(annotation_path)
+        y = int(line.split(';')[0])
+
+        pred        = classfication.detect_image(x)
+        pred_1      = np.argmax(pred)
+        correct_1   += pred_1 == y
+        
+        pred_5      = np.argsort(pred)[::-1]
+        pred_5      = pred_5[:5]
+        correct_5   += y in pred_5
+        
+        preds.append(pred_1)
+        labels.append(y)
+        if index % 100 == 0:
+            print("[%d/%d]"%(index, total))
+            
+    hist        = fast_hist(np.array(labels), np.array(preds), len(classfication.class_names))
+    Recall      = per_class_Recall(hist)
+    Precision   = per_class_Precision(hist)
+    
+    show_results(metrics_out_path, hist, Recall, Precision, classfication.class_names)
+    return correct_1 / total, correct_5 / total, Recall, Precision
+
+def fast_hist(a, b, n):
+    k = (a >= 0) & (a < n)
+    return np.bincount(n * a[k].astype(int) + b[k], minlength=n ** 2).reshape(n, n)  
+
+def per_class_Recall(hist):
+    return np.diag(hist) / np.maximum(hist.sum(1), 1) 
+
+def per_class_Precision(hist):
+    return np.diag(hist) / np.maximum(hist.sum(0), 1) 
+
+def adjust_axes(r, t, fig, axes):
+    bb                  = t.get_window_extent(renderer=r)
+    text_width_inches   = bb.width / fig.dpi
+    current_fig_width   = fig.get_figwidth()
+    new_fig_width       = current_fig_width + text_width_inches
+    propotion           = new_fig_width / current_fig_width
+    x_lim               = axes.get_xlim()
+    axes.set_xlim([x_lim[0], x_lim[1] * propotion])
+
+def draw_plot_func(values, name_classes, plot_title, x_label, output_path, tick_font_size = 12, plt_show = True):
+    fig     = plt.gcf() 
+    axes    = plt.gca()
+    plt.barh(range(len(values)), values, color='royalblue')
+    plt.title(plot_title, fontsize=tick_font_size + 2)
+    plt.xlabel(x_label, fontsize=tick_font_size)
+    plt.yticks(range(len(values)), name_classes, fontsize=tick_font_size)
+    r = fig.canvas.get_renderer()
+    for i, val in enumerate(values):
+        str_val = " " + str(val) 
+        if val < 1.0:
+            str_val = " {0:.2f}".format(val)
+        t = plt.text(val, i, str_val, color='royalblue', va='center', fontweight='bold')
+        if i == (len(values)-1):
+            adjust_axes(r, t, fig, axes)
+
+    fig.tight_layout()
+    fig.savefig(output_path)
+    if plt_show:
+        plt.show()
+    plt.close()
+    
+def show_results(miou_out_path, hist, Recall, Precision, name_classes, tick_font_size = 12):
+    draw_plot_func(Recall, name_classes, "mRecall = {0:.2f}%".format(np.nanmean(Recall)*100), "Recall", \
+        os.path.join(miou_out_path, "Recall.png"), tick_font_size = tick_font_size, plt_show = False)
+    print("Save Recall out to " + os.path.join(miou_out_path, "Recall.png"))
+
+    draw_plot_func(Precision, name_classes, "mPrecision = {0:.2f}%".format(np.nanmean(Precision)*100), "Precision", \
+        os.path.join(miou_out_path, "Precision.png"), tick_font_size = tick_font_size, plt_show = False)
+    print("Save Precision out to " + os.path.join(miou_out_path, "Precision.png"))
+
+    with open(os.path.join(miou_out_path, "confusion_matrix.csv"), 'w', newline='') as f:
+        writer          = csv.writer(f)
+        writer_list     = []
+        writer_list.append([' '] + [str(c) for c in name_classes])
+        for i in range(len(hist)):
+            writer_list.append([name_classes[i]] + [str(x) for x in hist[i]])
+        writer.writerows(writer_list)
+    print("Save confusion_matrix out to " + os.path.join(miou_out_path, "confusion_matrix.csv"))
+            
+def evaluteRecall(classfication, lines, metrics_out_path):
+    correct = 0
+    total = len(lines)
+    
+    preds   = []
+    labels  = []
+    for index, line in enumerate(lines):
+        annotation_path = line.split(';')[1].split()[0]
+        x = Image.open(annotation_path)
+        y = int(line.split(';')[0])
+
+        pred = classfication.detect_image(x)
+        pred = np.argmax(pred)
+        
+        preds.append(pred)
+        labels.append(y)
+        
+    hist        = fast_hist(labels, preds, len(classfication.class_names))
+    Recall      = per_class_Recall(hist)
+    Precision   = per_class_Precision(hist)
+    
+    show_results(metrics_out_path, hist, Recall, Precision, classfication.class_names)
+    return correct / total