├── README.md ├── VGG11.py ├── 数据集来源 ├── 测试集准确率数据集.csv ├── 测试集损失函数数据集.csv ├── 训练集准确率数据集.csv └── 训练集损失函数数据集.csv /README.md: -------------------------------------------------------------------------------- 1 | # Garbage-image-classification -------------------------------------------------------------------------------- /VGG11.py: -------------------------------------------------------------------------------- 1 | import matplotlib as mpl 2 | import matplotlib.pyplot as plt 3 | import numpy as np 4 | import pandas as pd 5 | import os 6 | import sys 7 | import time 8 | import tensorflow as tf 9 | from tensorflow import keras 10 | import sklearn 11 | from sklearn.preprocessing import StandardScaler 12 | import pathlib 13 | import random 14 | from PIL import Image 15 | from sklearn.model_selection import train_test_split 16 | 17 | data_path = pathlib.Path('./data/trash') 18 | all_image_paths = list(data_path.glob('*/*')) 19 | all_image_paths = [str(path) for path in all_image_paths] 20 | random.shuffle(all_image_paths) 21 | 22 | image_count = len(all_image_paths) 23 | label_names = sorted(item.name for item in data_path.glob('*/') if item.is_dir()) 24 | label_to_index = dict((name, index) for index, name in enumerate(label_names)) 25 | 26 | labels = [label_to_index[pathlib.Path(path).parent.name] for path in all_image_paths] 27 | images = [] 28 | 29 | def read_images(img_name): 30 | im = Image.open(img_name) 31 | data = np.array(im) 32 | data = tf.image.convert_image_dtype(data, dtype=tf.float32) #下面的色彩调整，必须是实数类型 33 | data = tf.image.random_flip_up_down(data) #随机上下翻转 34 | data = tf.image.random_flip_left_right(data) #随机左右翻转 35 | data = tf.image.random_brightness(data,max_delta=0.02) #色彩亮度调整，在(-max_delta, max_delta)的范围随机调整图像的亮度,0.1太大了，调整为0.02尝试 36 | data = tf.image.random_saturation(data, lower=0.9, upper=1.1) #饱和度调整 37 | data = tf.image.random_hue(data, max_delta=0.1) #调整色相 38 | data = tf.image.random_contrast(data, lower=0.9, upper=1.1) #调整对比度 39 | data = tf.clip_by_value(data, 0.0, 1.0) #最后对色彩进行调整，值限定再0-1之间,float32的格式， 40 | return np.array(data) 41 | 42 | 43 | def read_imagesnormal(img_name): 44 | im = Image.open(img_name) 45 | data = np.array(im) 46 | data = tf.image.convert_image_dtype(data, dtype=tf.float32) #下面的色彩调整，必须是实数类型 47 | data = tf.clip_by_value(data, 0.0, 1.0) #最后对色彩进行调整，值限定再0-1之间,float32的格式， 48 | return np.array(data) 49 | 50 | for img_name in all_image_paths[:1769]: #只需要对训练集进行数据预处理 51 | images.append(read_images(img_name)) 52 | for img_name in all_image_paths[1769:]: 53 | images.append(read_imagesnormal(img_name)) 54 | x=np.array(images) 55 | y = np.array(labels) 56 | 57 | x_train=x[:1769] 58 | y_train=y[:1769] 59 | x_test=x[1769:] 60 | y_test=y[1769:] 61 | 62 | scaler = StandardScaler() 63 | 64 | x_train_scaled = scaler.fit_transform(x_train.astype(np.float32).reshape(-1,1)).reshape(-1,384,512,3) #最后一个参数是通道数目 65 | x_test_sacled = scaler.transform(x_test.astype(np.float32).reshape(-1,1)).reshape(-1,384,512,3) 66 | 67 | 68 | 69 | def create_moedel(): 70 | model = keras.models.Sequential() 71 | model.add(keras.layers.Conv2D(filters=64,kernel_size=3,padding='same',activation='relu',input_shape=(384,512,3))) #添加卷积层操作，最后一个参数是通道数目 72 | model.add(keras.layers.MaxPool2D(pool_size=2,strides=2)) 73 | 74 | model.add(keras.layers.Conv2D(filters=128,kernel_size=3,padding='same',activation='relu')) 75 | model.add(keras.layers.MaxPool2D(pool_size=2,strides=2)) 76 | 77 | model.add(keras.layers.Conv2D(filters=256,kernel_size=3,padding='same',activation='relu')) 78 | model.add(keras.layers.Conv2D(filters=256,kernel_size=3,padding='same',activation='relu')) 79 | model.add(keras.layers.MaxPool2D(pool_size=2,strides=2)) 80 | 81 | model.add(keras.layers.Conv2D(filters=512,kernel_size=3,padding='same',activation='relu')) 82 | model.add(keras.layers.Conv2D(filters=512,kernel_size=3,padding='same',activation='relu')) 83 | model.add(keras.layers.MaxPool2D(pool_size=2,strides=2)) 84 | 85 | model.add(keras.layers.Conv2D(filters=512,kernel_size=3,padding='same',activation='relu')) 86 | model.add(keras.layers.Conv2D(filters=512,kernel_size=3,padding='same',activation='relu')) 87 | model.add(keras.layers.MaxPool2D(pool_size=2,strides=2)) 88 | 89 | model.add(keras.layers.Flatten()) 90 | model.add(keras.layers.Dropout(rate=0.5)) 91 | model.add(keras.layers.Dense(1024,kernel_regularizer=keras.regularizers.l2(0.05),activation='relu')) #L2正则化参数调整到0.05尝试 92 | model.add(keras.layers.Dropout(rate=0.5)) 93 | model.add(keras.layers.Dense(1024,kernel_regularizer=keras.regularizers.l2(0.05),activation='relu')) 94 | model.add(keras.layers.Dense(6,activation='softmax')) 95 | model.compile(loss="sparse_categorical_crossentropy",optimizer=tf.keras.optimizers.SGD(learning_rate=0.003),metrics =['accuracy']) #学习率也可以再调低一些0.004 96 | return model 97 | tb = tf.keras.callbacks.TensorBoard(log_dir='logs', # log 目录 98 | histogram_freq=5, # 按照何等频率（epoch）来计算直方图，0为不计算 99 | batch_size=40, # 用多大量的数据计算直方图 100 | write_graph=True, # 是否存储网络结构图 101 | write_grads=False, # 是否可视化梯度直方图 102 | write_images=True,# 是否可视化参数 103 | embeddings_freq=0, 104 | embeddings_layer_names=None, 105 | embeddings_metadata=None) 106 | model = create_moedel() 107 | print(model.summary()) 108 | early_stop = keras.callbacks.EarlyStopping(monitor='val_loss',patience=4) #连续10个epoch 验证集loss没有降低将停止训练 109 | history = model.fit(x_train_scaled,y_train,batch_size=20 ,epochs=100,validation_data=(x_test_sacled,y_test),callbacks=[early_stop,tb]) #一次给神经网络10个样本，批处理， 110 | print(history.history) 111 | def plot_learning_curves(history): 112 | pd.DataFrame(history.history).plot(figsize=(8,5)) 113 | plt.grid(True) #显示网格 114 | plt.gca().set_ylim(0,1) 115 | plt.show() 116 | plot_learning_curves(history) 117 | 118 | 119 | 120 | '''保存模型 121 | model.save('my_model.h5') 122 | new_model=keras.models.load_model('my_model.h5') 123 | new_model.summary() 124 | loss,acc=new_model.evaluate(x_test_sacled,y_test) 125 | print("Model accuracy:"+acc) 126 | ''' 127 | 128 | '''保存权重 129 | 130 | model.save_weights('./checkpoints/my_checkpoint') 131 | new2_model=create_model() 132 | new2_model.load_weights('./checkpoints/my_checkpoint') 133 | loss,acc=new2_model.evaluate(x_test_sacled,y_test) 134 | print("Model accuracy:"+acc) 135 | ''' 136 | 137 | 138 | ''' 139 | checkpoint_path="training_1/cp-1.ckpt" 140 | checkpoint_dir=os.path.dirname(checkpoint_path) 141 | cp_callback=tf.keras.callbacks.ModelCheckpoint(checkpoint_path,save_weight_only=True,verbose=1,period=10) #10次记录一次 142 | new3_model.fit(callbacks=[cp_callback]) #训练时放进入 143 | 144 | latest=tf.train.latest_checkpoint(checkpoint_dir) 145 | new4_model.load_weights(latest) #放进新模型 146 | 147 | ''' 148 | -------------------------------------------------------------------------------- /数据集来源: -------------------------------------------------------------------------------- 1 | https://github.com/garythung/trashnet 2 | -------------------------------------------------------------------------------- /测试集准确率数据集.csv: -------------------------------------------------------------------------------- 1 | Wall time,Step,Value 2 | 1591800597.100541,0,0.20976252853870392 3 | 1591800694.543333,1,0.20976252853870392 4 | 1591800784.644291,2,0.240105539560318 5 | 1591800874.395439,3,0.3153034448623657 6 | 1591800964.83789,4,0.31926122307777405 7 | 1591801056.005641,5,0.35092347860336304 8 | 1591801154.989915,6,0.3482849597930908 9 | 1591801248.141415,7,0.3799472153186798 10 | 1591801341.159041,8,0.3812665045261383 11 | 1591801434.668002,9,0.4155672788619995 12 | 1591801527.75274,10,0.3931398391723633 13 | 1591801628.237603,11,0.38786280155181885 14 | 1591801721.138556,12,0.4445910155773163 15 | 1591801812.90574,13,0.4472295641899109 16 | 1591801908.523823,14,0.47361478209495544 17 | 1591802002.641839,15,0.47757256031036377 18 | 1591802100.195633,16,0.468337744474411 19 | 1591802189.826882,17,0.5026385188102722 20 | 1591802279.630669,18,0.4854881167411804 21 | 1591802369.695782,19,0.5 22 | 1591802459.92187,20,0.5211081504821777 23 | 1591802558.273079,21,0.49208444356918335 24 | 1591802648.265193,22,0.5039578080177307 25 | 1591802738.766409,23,0.5488126873970032 26 | 1591802830.576597,24,0.5540897250175476 27 | 1591802920.835243,25,0.5527704358100891 28 | 1591803019.528319,26,0.539577841758728 29 | 1591803111.029203,27,0.5184696316719055 30 | 1591803203.613694,28,0.4445910155773163 31 | 1591803295.072842,29,0.5738786458969116 32 | 1591803387.71604,30,0.49076518416404724 33 | 1591803486.050546,31,0.5422163605690002 34 | 1591803576.176406,32,0.5751978754997253 35 | 1591803669.768829,33,0.5026385188102722 36 | 1591803761.644875,34,0.5712401270866394 37 | 1591803854.454639,35,0.4986807405948639 38 | 1591803954.546967,36,0.565963089466095 39 | 1591804047.139832,37,0.565963089466095 40 | 1591804138.324291,38,0.5237467288970947 41 | 1591804228.234105,39,0.5672823190689087 42 | 1591804319.192902,40,0.5738786458969116 43 | 1591804417.811184,41,0.5976253151893616 44 | 1591804508.637911,42,0.5725593566894531 45 | 1591804599.571209,43,0.5778363943099976 46 | 1591804691.422651,44,0.6042216420173645 47 | 1591804782.123227,45,0.5646438002586365 48 | 1591804880.824865,46,0.5672823190689087 49 | 1591804970.600548,47,0.5844327211380005 50 | 1591805059.785147,48,0.5501319169998169 51 | 1591805148.969143,49,0.5646438002586365 52 | 1591805238.195353,50,0.5131925940513611 53 | 1591805336.38793,51,0.5949867963790894 54 | 1591805425.572549,52,0.5158311128616333 55 | 1591805514.756704,53,0.5857519507408142 56 | 1591805603.957641,54,0.6042216420173645 57 | 1591805693.125721,55,0.5686016082763672 58 | 1591805790.651798,56,0.5065963268280029 59 | 1591805879.844593,57,0.5646438002586365 60 | 1591805969.02011,58,0.5290237665176392 61 | 1591806058.313179,59,0.5712401270866394 62 | 1591806147.622707,60,0.5844327211380005 63 | 1591806245.006572,61,0.539577841758728 64 | 1591806334.141334,62,0.5870712399482727 65 | 1591806423.350619,63,0.5883905291557312 66 | 1591806512.968164,64,0.5540897250175476 67 | 1591806606.694266,65,0.5844327211380005 68 | 1591806706.970342,66,0.5765171647071838 69 | 1591806797.896087,67,0.5606859922409058 70 | 1591806887.797258,68,0.6002638339996338 71 | 1591806977.098609,69,0.583113431930542 72 | 1591807066.615755,70,0.5514512062072754 73 | 1591807166.092295,71,0.583113431930542 74 | 1591807257.426787,72,0.6055408716201782 75 | 1591807349.894168,73,0.5804749131202698 76 | 1591807440.295618,74,0.602902352809906 77 | 1591807530.162896,75,0.5976253151893616 78 | 1591807629.972543,76,0.5105540752410889 79 | 1591807723.290471,77,0.5382585525512695 80 | 1591807814.482563,78,0.5976253151893616 81 | 1591807908.409203,79,0.6094986796379089 82 | 1591808002.468313,80,0.4696570038795471 83 | 1591808104.211381,81,0.5712401270866394 84 | 1591808194.678641,82,0.4709762632846832 85 | 1591808286.129832,83,0.6015831232070923 86 | 1591808377.147476,84,0.579155683517456 87 | 1591808467.556667,85,0.6002638339996338 88 | 1591808566.341251,86,0.5883905291557312 89 | 1591808657.23427,87,0.6002638339996338 90 | 1591808748.301712,88,0.6121371984481812 91 | 1591808838.102715,89,0.5725593566894531 92 | 1591808927.56249,90,0.5897097587585449 93 | 1591809025.854863,91,0.5910290479660034 94 | 1591809115.289627,92,0.563324511051178 95 | 1591809204.67357,93,0.5422163605690002 96 | 1591809294.250033,94,0.5738786458969116 97 | 1591809383.917886,95,0.563324511051178 98 | 1591809482.426953,96,0.5804749131202698 99 | 1591809572.586284,97,0.5501319169998169 100 | 1591809664.47867,98,0.565963089466095 101 | 1591809757.738215,99,0.579155683517456 102 | -------------------------------------------------------------------------------- /测试集损失函数数据集.csv: -------------------------------------------------------------------------------- 1 | Wall time,Step,Value 2 | 1591800597.100394,0,146.3745880126953 3 | 1591800694.543179,1,138.8424072265625 4 | 1591800784.644086,2,131.66773986816406 5 | 1591800874.395286,3,124.8379135131836 6 | 1591800964.837747,4,118.4317855834961 7 | 1591801056.005475,5,112.33343505859375 8 | 1591801154.989685,6,106.56513214111328 9 | 1591801248.141251,7,101.0691146850586 10 | 1591801341.158897,8,95.87570190429688 11 | 1591801434.667853,9,90.93621063232422 12 | 1591801527.752523,10,86.36817169189453 13 | 1591801628.237453,11,81.96884155273438 14 | 1591801721.1384,12,77.65151977539062 15 | 1591801812.905582,13,73.69913482666016 16 | 1591801908.523664,14,69.88301849365234 17 | 1591802002.641684,15,66.3205337524414 18 | 1591802100.195401,16,62.97285461425781 19 | 1591802189.826729,17,59.70020294189453 20 | 1591802279.630487,18,56.693729400634766 21 | 1591802369.695592,19,53.76520538330078 22 | 1591802459.921715,20,51.05471420288086 23 | 1591802558.272914,21,48.52098846435547 24 | 1591802648.265043,22,46.04361343383789 25 | 1591802738.766259,23,43.6635627746582 26 | 1591802830.5764,24,41.47454833984375 27 | 1591802920.835087,25,39.354915618896484 28 | 1591803019.528162,26,37.439292907714844 29 | 1591803111.029026,27,35.69497299194336 30 | 1591803203.613528,28,34.05499267578125 31 | 1591803295.072697,29,32.01995849609375 32 | 1591803387.715854,30,30.665964126586914 33 | 1591803486.050387,31,29.012895584106445 34 | 1591803576.176255,32,27.552326202392578 35 | 1591803669.768677,33,26.41073226928711 36 | 1591803761.644689,34,24.883647918701172 37 | 1591803854.454478,35,23.906116485595703 38 | 1591803954.546813,36,22.53679084777832 39 | 1591804047.139668,37,21.555444717407227 40 | 1591804138.324142,38,20.566314697265625 41 | 1591804228.233937,39,19.480066299438477 42 | 1591804319.192717,40,18.606956481933594 43 | 1591804417.811027,41,17.588367462158203 44 | 1591804508.63758,42,16.822154998779297 45 | 1591804599.571054,43,16.109846115112305 46 | 1591804691.422502,44,15.27193832397461 47 | 1591804782.123077,45,14.819657325744629 48 | 1591804880.824715,46,14.182180404663086 49 | 1591804970.600359,47,13.524147033691406 50 | 1591805059.785023,48,13.05070972442627 51 | 1591805148.968999,49,12.38959789276123 52 | 1591805238.19521,50,12.353053092956543 53 | 1591805336.387763,51,11.40234661102295 54 | 1591805425.5724,52,11.29886531829834 55 | 1591805514.756483,53,10.473135948181152 56 | 1591805603.957485,54,10.157341003417969 57 | 1591805693.125571,55,10.045247077941895 58 | 1591805790.651652,56,9.760257720947266 59 | 1591805879.844443,57,9.086237907409668 60 | 1591805969.019954,58,9.513895034790039 61 | 1591806058.313007,59,8.449028015136719 62 | 1591806147.622504,60,8.164557456970215 63 | 1591806245.006418,61,8.060053825378418 64 | 1591806334.141182,62,7.487237453460693 65 | 1591806423.350461,63,7.642822742462158 66 | 1591806512.968033,64,7.344684600830078 67 | 1591806606.694112,65,6.889413356781006 68 | 1591806706.970196,66,6.7533111572265625 69 | 1591806797.895939,67,7.0782365798950195 70 | 1591806887.797095,68,6.109066963195801 71 | 1591806977.098454,69,6.193660259246826 72 | 1591807066.61562,70,6.166027069091797 73 | 1591807166.092148,71,5.621081352233887 74 | 1591807257.426594,72,5.627162456512451 75 | 1591807349.894019,73,5.5600996017456055 76 | 1591807440.295461,74,5.038601398468018 77 | 1591807530.162738,75,5.183874607086182 78 | 1591807629.972395,76,6.2121262550354 79 | 1591807723.290271,77,5.587218761444092 80 | 1591807814.482405,78,4.622154235839844 81 | 1591807908.409045,79,4.576513767242432 82 | 1591808002.468115,80,6.579917907714844 83 | 1591808104.211225,81,4.4696574211120605 84 | 1591808194.678487,82,5.080073356628418 85 | 1591808286.129643,83,4.278668403625488 86 | 1591808377.147318,84,4.3011298179626465 87 | 1591808467.556529,85,4.029945373535156 88 | 1591808566.341101,86,3.996486186981201 89 | 1591808657.234125,87,4.095329761505127 90 | 1591808748.301543,88,3.7529401779174805 91 | 1591808838.102562,89,3.9933183193206787 92 | 1591808927.562346,90,3.998283624649048 93 | 1591809025.854706,91,3.589846134185791 94 | 1591809115.289471,92,3.8727433681488037 95 | 1591809204.673424,93,3.757420063018799 96 | 1591809294.249824,94,3.868364095687866 97 | 1591809383.917709,95,3.9395978450775146 98 | 1591809482.42681,96,3.3042445182800293 99 | 1591809572.586043,97,3.4793295860290527 100 | 1591809664.478525,98,3.363776445388794 101 | 1591809757.73803,99,3.25892972946167 102 | -------------------------------------------------------------------------------- /训练集准确率数据集.csv: -------------------------------------------------------------------------------- 1 | Wall time,Step,Value 2 | 1591800596.770369,0,0.22272470593452454 3 | 1591800694.253491,1,0.2453363537788391 4 | 1591800784.588328,2,0.2549462914466858 5 | 1591800874.111278,3,0.2928208112716675 6 | 1591800964.55054,4,0.33860939741134644 7 | 1591801055.946655,5,0.3431317210197449 8 | 1591801154.71004,6,0.37987563014030457 9 | 1591801247.857595,7,0.37874504923820496 10 | 1591801340.879834,8,0.42792537808418274 11 | 1591801434.389491,9,0.42792537808418274 12 | 1591801527.463705,10,0.4460147023200989 13 | 1591801627.946113,11,0.4511023163795471 14 | 1591801720.843611,12,0.4759751260280609 15 | 1591801812.620332,13,0.484454482793808 16 | 1591801907.931032,14,0.49462974071502686 17 | 1591802002.355789,15,0.5070661306381226 18 | 1591802100.131592,16,0.5223289728164673 19 | 1591802189.541199,17,0.5268513560295105 20 | 1591802279.347508,18,0.5415489077568054 21 | 1591802369.409386,19,0.5528547167778015 22 | 1591802459.636124,20,0.5545505881309509 23 | 1591802557.994593,21,0.5692481398582458 24 | 1591802647.984607,22,0.5788581371307373 25 | 1591802738.720681,23,0.5635952353477478 26 | 1591802830.284915,24,0.5924251079559326 27 | 1591802920.552062,25,0.6042962074279785 28 | 1591803019.248494,26,0.6054267883300781 29 | 1591803110.747558,27,0.6127755641937256 30 | 1591803203.334812,28,0.6167325973510742 31 | 1591803294.790113,29,0.6257773041725159 32 | 1591803387.439712,30,0.6438665986061096 33 | 1591803485.765572,31,0.6546071171760559 34 | 1591803575.888,32,0.6534765362739563 35 | 1591803669.482813,33,0.6636517643928528 36 | 1591803761.352361,34,0.6794799566268921 37 | 1591803854.165659,35,0.6890898942947388 38 | 1591803954.25677,36,0.6981345415115356 39 | 1591804047.089663,37,0.7179197072982788 40 | 1591804138.038369,38,0.7105709314346313 41 | 1591804227.943238,39,0.7507066130638123 42 | 1591804318.903264,40,0.7524024844169617 43 | 1591804417.531911,41,0.7377049326896667 44 | 1591804508.350197,42,0.7755794525146484 45 | 1591804599.294893,43,0.7772753238677979 46 | 1591804691.138473,44,0.7727529406547546 47 | 1591804781.841627,45,0.8106274604797363 48 | 1591804880.544851,46,0.8157150745391846 49 | 1591804970.312838,47,0.8349350094795227 50 | 1591805059.50427,48,0.84284907579422 51 | 1591805148.68651,49,0.8530243039131165 52 | 1591805237.915308,50,0.8733747601509094 53 | 1591805336.110292,51,0.8920294046401978 54 | 1591805425.290275,52,0.8993781805038452 55 | 1591805514.478007,53,0.902769923210144 56 | 1591805603.675855,54,0.9191633462905884 57 | 1591805692.842237,55,0.9135104417800903 58 | 1591805790.36456,56,0.8886376619338989 59 | 1591805879.552723,57,0.931034505367279 60 | 1591805968.732875,58,0.9406444430351257 61 | 1591806058.02094,59,0.931034505367279 62 | 1591806147.333878,60,0.9434708952903748 63 | 1591806244.713989,61,0.9412097334861755 64 | 1591806333.85053,62,0.9265121817588806 65 | 1591806423.058469,63,0.9666478037834167 66 | 1591806512.677836,64,0.9570378661155701 67 | 1591806606.405661,65,0.9683436751365662 68 | 1591806706.676065,66,0.9609948992729187 69 | 1591806797.806826,67,0.968908965587616 70 | 1591806887.505249,68,0.9474279284477234 71 | 1591806976.807107,69,0.9677783846855164 72 | 1591807066.322848,70,0.9638213515281677 73 | 1591807165.805204,71,0.9706048369407654 74 | 1591807257.380058,72,0.9836065769195557 75 | 1591807349.602552,73,0.9626907706260681 76 | 1591807440.002793,74,0.9632560610771179 77 | 1591807529.872127,75,0.976823091506958 78 | 1591807629.679599,76,0.974561870098114 79 | 1591807723.003778,77,0.968908965587616 80 | 1591807814.189974,78,0.9530808329582214 81 | 1591807908.117186,79,0.9785189628601074 82 | 1591808002.174776,80,0.976823091506958 83 | 1591808103.917333,81,0.9615601897239685 84 | 1591808194.387968,82,0.9773883819580078 85 | 1591808285.841091,83,0.9773883819580078 86 | 1591808376.859557,84,0.9898247718811035 87 | 1591808467.266391,85,0.9734312891960144 88 | 1591808566.030674,86,0.9694742560386658 89 | 1591808656.935998,87,0.9807801246643066 90 | 1591808748.012321,88,0.982475996017456 91 | 1591808837.81152,89,0.9819107055664062 92 | 1591808927.270908,90,0.982475996017456 93 | 1591809025.561581,91,0.945732057094574 94 | 1591809114.996609,92,0.9847371578216553 95 | 1591809204.379243,93,0.9819107055664062 96 | 1591809293.956585,94,0.9903900623321533 97 | 1591809383.628462,95,0.9915206432342529 98 | 1591809482.135019,96,0.9576031565666199 99 | 1591809572.556528,97,0.9728659987449646 100 | 1591809664.423949,98,0.9751272201538086 101 | 1591809757.446137,99,0.9728659987449646 102 | -------------------------------------------------------------------------------- /训练集损失函数数据集.csv: -------------------------------------------------------------------------------- 1 | Wall time,Step,Value 2 | 1591800596.770206,0,150.40553283691406 3 | 1591800694.253351,1,142.62229919433594 4 | 1591800784.588188,2,135.2761688232422 5 | 1591800874.111141,3,128.28936767578125 6 | 1591800964.550186,4,121.66007232666016 7 | 1591801055.946503,5,115.40026092529297 8 | 1591801154.709888,6,109.4502182006836 9 | 1591801247.857467,7,103.82907104492188 10 | 1591801340.879688,8,98.47441864013672 11 | 1591801434.38936,9,93.4111557006836 12 | 1591801527.463571,10,88.62178802490234 13 | 1591801627.945974,11,84.06877899169922 14 | 1591801720.84347,12,79.74665832519531 15 | 1591801812.620191,13,75.6572036743164 16 | 1591801907.930884,14,71.76493072509766 17 | 1591802002.355643,15,68.1017837524414 18 | 1591802100.131455,16,64.60738372802734 19 | 1591802189.541064,17,61.29096603393555 20 | 1591802279.347372,18,58.16288757324219 21 | 1591802369.40924,19,55.195125579833984 22 | 1591802459.635979,20,52.36345291137695 23 | 1591802557.994452,21,49.67237854003906 24 | 1591802647.984479,22,47.15126037597656 25 | 1591802738.720535,23,44.77926254272461 26 | 1591802830.284775,24,42.4642333984375 27 | 1591802920.551922,25,40.30356216430664 28 | 1591803019.248369,26,38.25986862182617 29 | 1591803110.74742,27,36.325565338134766 30 | 1591803203.334672,28,34.47591018676758 31 | 1591803294.789984,29,32.748146057128906 32 | 1591803387.439571,30,31.074813842773438 33 | 1591803485.76543,31,29.520658493041992 34 | 1591803575.887862,32,28.02751922607422 35 | 1591803669.482687,33,26.610078811645508 36 | 1591803761.352219,34,25.269947052001953 37 | 1591803854.165485,35,23.97873306274414 38 | 1591803954.256624,36,22.804061889648438 39 | 1591804047.08952,37,21.625263214111328 40 | 1591804138.038238,38,20.544485092163086 41 | 1591804227.943098,39,19.477983474731445 42 | 1591804318.903092,40,18.522254943847656 43 | 1591804417.531751,41,17.642066955566406 44 | 1591804508.350047,42,16.69930076599121 45 | 1591804599.294757,43,15.868440628051758 46 | 1591804691.138342,44,15.0973539352417 47 | 1591804781.841487,45,14.304781913757324 48 | 1591804880.544693,46,13.62746524810791 49 | 1591804970.3127,47,12.926065444946289 50 | 1591805059.504133,48,12.286184310913086 51 | 1591805148.686372,49,11.660674095153809 52 | 1591805237.915161,50,11.101151466369629 53 | 1591805336.110156,51,10.523959159851074 54 | 1591805425.290125,52,9.999064445495605 55 | 1591805514.477864,53,9.544181823730469 56 | 1591805603.675709,54,9.049912452697754 57 | 1591805692.842097,55,8.640227317810059 58 | 1591805790.364416,56,8.350488662719727 59 | 1591805879.552571,57,7.8500542640686035 60 | 1591805968.732735,58,7.456173896789551 61 | 1591806058.020809,59,7.142367839813232 62 | 1591806147.333727,60,6.7834649085998535 63 | 1591806244.713858,61,6.488160133361816 64 | 1591806333.850376,62,6.26693058013916 65 | 1591806423.058322,63,5.854227542877197 66 | 1591806512.677578,64,5.601464748382568 67 | 1591806606.405526,65,5.329033374786377 68 | 1591806706.675929,66,5.101381778717041 69 | 1591806797.806686,67,4.855216026306152 70 | 1591806887.505118,68,4.706089496612549 71 | 1591806976.806934,69,4.461367607116699 72 | 1591807066.322701,70,4.269044876098633 73 | 1591807165.805054,71,4.06892728805542 74 | 1591807257.37991,72,3.8421311378479004 75 | 1591807349.602404,73,3.7405142784118652 76 | 1591807440.00258,74,3.5751988887786865 77 | 1591807529.871947,75,3.3950724601745605 78 | 1591807629.679442,76,3.25242280960083 79 | 1591807723.003641,77,3.1338510513305664 80 | 1591807814.189838,78,3.07598614692688 81 | 1591807908.117043,79,2.8770101070404053 82 | 1591808002.174641,80,2.756166934967041 83 | 1591808103.917204,81,2.725959300994873 84 | 1591808194.387829,82,2.553187608718872 85 | 1591808285.840962,83,2.476417303085327 86 | 1591808376.859421,84,2.322136402130127 87 | 1591808467.266257,85,2.287355661392212 88 | 1591808566.03053,86,2.21934175491333 89 | 1591808656.93584,87,2.093677282333374 90 | 1591808748.012182,88,2.0154366493225098 91 | 1591808837.811381,89,1.9333784580230713 92 | 1591808927.270771,90,1.8795253038406372 93 | 1591809025.561441,91,1.961940884590149 94 | 1591809114.996468,92,1.7737774848937988 95 | 1591809204.379104,93,1.7147479057312012 96 | 1591809293.956441,94,1.623850703239441 97 | 1591809383.628337,95,1.5483015775680542 98 | 1591809482.134881,96,1.6375492811203003 99 | 1591809572.556382,97,1.5437557697296143 100 | 1591809664.423754,98,1.4869139194488525 101 | 1591809757.446009,99,1.4462391138076782 102 | --------------------------------------------------------------------------------