构建一个字母abc的手写识别网络,
要求给出算法误差收敛曲线,所给程序要有图片导入接口。
其中a,b,c都代表label,三个文件夹存在具体的图片。只要是这样类型的,直接套下面模板。
import os
import cv2
import numpy
as np
import pandas
as pd
import matplotlib
.pyplot
as plt
import tensorflow
as tf
import tqdm
from tensorflow
import keras
from keras
import input
, model
, sequential
from tensorflow
.keras
.regularizers
import l2
from keras
.layers
import dense
, flatten
, inputlayer
, reshape
, batchnormalization
, dropout
, conv2d
, maxpooling2d
from tensorflow
.keras
.utils
import plot_model
%matplotlib inline
data_dir
= './data'
categories
= {
'a': 0,
'b': 1,
'c': 2
}
def load_images(images_folder
, img_size
= (128,128), scale
=false):image_path
= []for dirname
, _
, filenames
in os
.walk
(images_folder
):for filename
in filenames
:image_path
.append
(os
.path
.join
(dirname
, filename
))print("there are {} images in {}".format(len(image_path
), images_folder
))images
= []labels
= []for path
in tqdm
.tqdm
(image_path
):img
= cv2
.imread
(path
) img
= cv2
.resize
(img
, img_size
) img
= np
.array
(img
)images
.append
(img
)labels
.append
(categories
[path
.split
('/')[-2]]) images
= np
.array
(images
) images
= images
.astype
(np
.int64
)if scale
:images
= images
/255 return image_path
, images
, np
.asarray
(labels
)
img_size
= (128,128)
image_path
, images
, labels
= load_images
(data_dir
, img_size
=img_size
)
images
.shape
there are 600 images in ./data100%|██████████| 600/600 [00:03<00:00, 183.15it/s](600, 128, 128, 3)
plt
.figure
(figsize
=(10,10))
random_inds
= np
.random
.choice
(len(image_path
),36)
for i
in range(36):plt
.subplot
(6,6,i
1)plt
.xticks
([])plt
.yticks
([])plt
.grid
(false)image_ind
= random_inds
[i
]plt
.imshow
(np
.squeeze
(images
[image_ind
]), cmap
=plt
.cm
.binary
)label
= list(categories
.keys
())[list(categories
.values
()).index
(labels
[image_ind
])]plt
.title
(label
)
labels_df
= pd
.dataframe
(labels
)
labels_df
.value_counts
()
2 201
0 201
1 198
dtype: int64
dataset
=[]
dataname
=[]
count
=0
for name
in tqdm
(os
.listdir
(data_dir
)):path
=os
.path
.join
(data_dir
,name
)for im
in os
.listdir
(path
):image
=cv2
.imread
(os
.path
.join
(path
,im
))image2
=np
.resize
(image
,(50,50,3))dataset
=[image2
]dataname
=[count
]count
=count
1
100%|██████████| 3/3 [00:03<00:00, 1.06s/it]
data
=np
.array
(dataset
)
dataname
=np
.array
(dataname
)
data
[0].shape
(50, 50, 3)
print(pd
.series
(dataname
).value_counts
())
1 202
2 201
0 198
dtype: int64
len(categories
)
3
from tensorflow
.keras
.layers
import dense
, conv2d
, flatten
, dropout
, maxpooling2d
, batchnormalization
def build_cnn_model():cnn_model
=tf
.keras
.sequential
([conv2d
(filters
=32,kernel_size
=(3,3),activation
='relu',input_shape
=images
.shape
[1:]),maxpooling2d
(2,2),batchnormalization
(),dropout
(0.4),conv2d
(filters
=64,kernel_size
=(3,3),activation
='relu', padding
='same'),conv2d
(filters
=64,kernel_size
=(3,3),activation
='relu', padding
='same'),maxpooling2d
((2,2)),batchnormalization
(),dropout
(0.4),conv2d
(filters
=128,kernel_size
=(3,3),activation
='relu', padding
='same'),conv2d
(filters
=128,kernel_size
=(3,3),activation
='relu', padding
='same'),maxpooling2d
(2,2),batchnormalization
(),dropout
(0.4),conv2d
(filters
=256,kernel_size
=(3,3),activation
='relu', padding
='same'),conv2d
(filters
=256,kernel_size
=(3,3),activation
='relu', padding
='same'),maxpooling2d
(2,2),batchnormalization
(),dropout
(0.4),conv2d
(filters
=128,kernel_size
=(3,3),activation
='relu', padding
='same'),conv2d
(filters
=128,kernel_size
=(3,3),activation
='relu', padding
='same'),maxpooling2d
(2,2),batchnormalization
(),dropout
(0.4),conv2d
(filters
=64,kernel_size
=(3,3),activation
='relu', padding
='same'),conv2d
(filters
=64,kernel_size
=(3,3),activation
='relu', padding
='same'),maxpooling2d
((2,2)),batchnormalization
(),dropout
(0.4),flatten
(),dense
(units
=len(categories
),activation
='softmax')])return cnn_modelmodel
= build_cnn_model
()
model
.predict
(images
[[0]])
print(model
.summary
())
model: "sequential_3"
_________________________________________________________________layer (type) output shape param #
=================================================================conv2d_6 (conv2d) (none, 126, 126, 32) 896 max_pooling2d_6 (maxpooling (none, 63, 63, 32) 0 2d) batch_normalization (batchn (none, 63, 63, 32) 128 ormalization) dropout (dropout) (none, 63, 63, 32) 0 conv2d_7 (conv2d) (none, 63, 63, 64) 18496 conv2d_8 (conv2d) (none, 63, 63, 64) 36928 max_pooling2d_7 (maxpooling (none, 31, 31, 64) 0 2d) batch_normalization_1 (batc (none, 31, 31, 64) 256 hnormalization) dropout_1 (dropout) (none, 31, 31, 64) 0 conv2d_9 (conv2d) (none, 31, 31, 128) 73856 conv2d_10 (conv2d) (none, 31, 31, 128) 147584 max_pooling2d_8 (maxpooling (none, 15, 15, 128) 0 2d) batch_normalization_2 (batc (none, 15, 15, 128) 512 hnormalization) dropout_2 (dropout) (none, 15, 15, 128) 0 conv2d_11 (conv2d) (none, 15, 15, 256) 295168 conv2d_12 (conv2d) (none, 15, 15, 256) 590080 max_pooling2d_9 (maxpooling (none, 7, 7, 256) 0 2d) batch_normalization_3 (batc (none, 7, 7, 256) 1024 hnormalization) dropout_3 (dropout) (none, 7, 7, 256) 0 conv2d_13 (conv2d) (none, 7, 7, 128) 295040 conv2d_14 (conv2d) (none, 7, 7, 128) 147584 max_pooling2d_10 (maxpoolin (none, 3, 3, 128) 0 g2d) batch_normalization_4 (batc (none, 3, 3, 128) 512 hnormalization) dropout_4 (dropout) (none, 3, 3, 128) 0 conv2d_15 (conv2d) (none, 3, 3, 64) 73792 conv2d_16 (conv2d) (none, 3, 3, 64) 36928 max_pooling2d_11 (maxpoolin (none, 1, 1, 64) 0 g2d) batch_normalization_5 (batc (none, 1, 1, 64) 256 hnormalization) dropout_5 (dropout) (none, 1, 1, 64) 0 flatten_1 (flatten) (none, 64) 0 dense_6 (dense) (none, 3) 195 =================================================================
total params: 1,719,235
trainable params: 1,717,891
non-trainable params: 1,344
_________________________________________________________________
none
tf
.keras
.utils
.plot_model
(model
, show_shapes
=true)
from tensorflow
.keras
.utils
import to_categorical
from sklearn
.preprocessing
import labelencoder
from sklearn
.utils
import shuffle
le
= labelencoder
()
labels
= le
.fit_transform
(labels
)
labels
= to_categorical
(labels
)
labels
[:10]
array([[0., 1., 0.],[0., 1., 0.],[0., 1., 0.],[0., 1., 0.],[0., 1., 0.],[0., 1., 0.],[0., 1., 0.],[0., 1., 0.],[0., 1., 0.],[0., 1., 0.]], dtype=float32)
model
.compile(optimizer
= "adam", loss
= "binary_crossentropy", metrics
= ["accuracy"])
checkpoint_filepath
= '/checkpoint.hdf5'
model_checkpoint_callback
= tf
.keras
.callbacks
.modelcheckpoint
(filepath
=checkpoint_filepath
,save_weights_only
=true,monitor
='val_accuracy',mode
='max',save_best_only
=true,save_freq
=500)
from tensorflow
.keras
.callbacks
import earlystopping
early_stopping
= earlystopping
(monitor
='val_accuracy', patience
=10, min_delta
=0.001, mode
='max',restore_best_weights
=true
)
datagen
= imagedatagenerator
(horizontal_flip
=true,vertical_flip
=true,rotation_range
=20,zoom_range
=0.2,width_shift_range
=0.2,height_shift_range
=0.2,shear_range
=0.1,fill_mode
="nearest")
from tensorflow
.keras
.callbacks
import reducelronplateau
reducelr
= reducelronplateau
(monitor
= "val_accuracy",factor
= 0.3, patience
= 3,min_delta
= 0.001,mode
= 'auto',verbose
=1)
from sklearn
.model_selection
import train_test_split
x_train
, x_test
, y_train
, y_test
= train_test_split
(images
, labels
, test_size
=0.10, random_state
=7)x_train
, x_val
, y_train
, y_val
= train_test_split
(x_train
, y_train
, test_size
=0.20, random_state
=1)
print("*-*-*-*-*-*")
print("train")
print(x_train
.shape
)
print(y_train
.shape
)print("*-*-*-*-*-*")
print("validation")
print(x_val
.shape
)
print(y_val
.shape
)print("*-*-*-*-*-*")
print("test")
print(x_test
.shape
)
print(y_test
.shape
)
*-*-*-*-*-*
train
(432, 128, 128, 3)
(432, 3)
*-*-*-*-*-*
validation
(108, 128, 128, 3)
(108, 3)
*-*-*-*-*-*
test
(60, 128, 128, 3)
(60, 3)
history
= model
.fit
(x_train
, y_train
, batch_size
= 32, epochs
= 100, verbose
= 1, validation_data
= (x_val
, y_val
),callbacks
=[model_checkpoint_callback
, early_stopping
, reducelr
])
epoch 1/1000
14/14 [==============================] - 8s 238ms/step - loss: 0.8036 - accuracy: 0.3588 - val_loss: 6.2921 - val_accuracy: 0.2963 - lr: 0.0010
epoch 2/1000
14/14 [==============================] - 1s 95ms/step - loss: 0.8162 - accuracy: 0.3796 - val_loss: 5.2361 - val_accuracy: 0.2963 - lr: 0.0010
epoch 3/1000
14/14 [==============================] - 1s 97ms/step - loss: 0.7190 - accuracy: 0.4537 - val_loss: 1.3893 - val_accuracy: 0.3333 - lr: 0.0010
epoch 4/1000
14/14 [==============================] - 1s 97ms/step - loss: 0.6875 - accuracy: 0.4792 - val_loss: 0.7386 - val_accuracy: 0.3519 - lr: 0.0010
epoch 5/1000
14/14 [==============================] - 1s 100ms/step - loss: 0.6144 - accuracy: 0.5949 - val_loss: 0.7014 - val_accuracy: 0.4259 - lr: 0.0010
epoch 6/1000
14/14 [==============================] - 1s 97ms/step - loss: 0.5156 - accuracy: 0.7060 - val_loss: 0.7592 - val_accuracy: 0.4537 - lr: 0.0010
epoch 7/1000
14/14 [==============================] - 1s 96ms/step - loss: 0.4904 - accuracy: 0.7384 - val_loss: 0.7034 - val_accuracy: 0.5370 - lr: 0.0010
epoch 8/1000
14/14 [==============================] - 1s 97ms/step - loss: 0.3854 - accuracy: 0.7940 - val_loss: 0.6092 - val_accuracy: 0.5556 - lr: 0.0010
epoch 9/1000
14/14 [==============================] - 1s 97ms/step - loss: 0.3313 - accuracy: 0.8241 - val_loss: 0.5192 - val_accuracy: 0.6389 - lr: 0.0010
epoch 10/1000
14/14 [==============================] - 1s 93ms/step - loss: 0.2873 - accuracy: 0.8519 - val_loss: 0.5089 - val_accuracy: 0.6111 - lr: 0.0010
epoch 11/1000
14/14 [==============================] - 1s 96ms/step - loss: 0.2346 - accuracy: 0.8981 - val_loss: 0.4359 - val_accuracy: 0.6852 - lr: 0.0010
epoch 12/1000
14/14 [==============================] - 1s 94ms/step - loss: 0.2238 - accuracy: 0.8819 - val_loss: 0.4404 - val_accuracy: 0.6481 - lr: 0.0010
epoch 13/1000
14/14 [==============================] - 1s 97ms/step - loss: 0.1954 - accuracy: 0.8912 - val_loss: 0.4215 - val_accuracy: 0.7500 - lr: 0.0010
epoch 14/1000
14/14 [==============================] - 1s 100ms/step - loss: 0.1792 - accuracy: 0.9051 - val_loss: 0.1971 - val_accuracy: 0.9074 - lr: 0.0010
epoch 15/1000
14/14 [==============================] - 1s 96ms/step - loss: 0.1608 - accuracy: 0.9144 - val_loss: 0.2836 - val_accuracy: 0.8056 - lr: 0.0010
epoch 16/1000
14/14 [==============================] - 1s 95ms/step - loss: 0.1447 - accuracy: 0.9398 - val_loss: 0.2867 - val_accuracy: 0.7500 - lr: 0.0010
epoch 17/1000
14/14 [==============================] - eta: 0s - loss: 0.1215 - accuracy: 0.9375
epoch 00017: reducelronplateau reducing learning rate to 0.0003000000142492354.
14/14 [==============================] - 1s 95ms/step - loss: 0.1215 - accuracy: 0.9375 - val_loss: 0.1474 - val_accuracy: 0.9074 - lr: 0.0010
epoch 18/1000
14/14 [==============================] - 1s 97ms/step - loss: 0.1023 - accuracy: 0.9537 - val_loss: 0.1186 - val_accuracy: 0.9352 - lr: 3.0000e-04
epoch 19/1000
14/14 [==============================] - 1s 101ms/step - loss: 0.0992 - accuracy: 0.9606 - val_loss: 0.1074 - val_accuracy: 0.9444 - lr: 3.0000e-04
epoch 20/1000
14/14 [==============================] - 1s 94ms/step - loss: 0.0837 - accuracy: 0.9676 - val_loss: 0.0917 - val_accuracy: 0.9444 - lr: 3.0000e-04
epoch 21/1000
14/14 [==============================] - 1s 98ms/step - loss: 0.0788 - accuracy: 0.9699 - val_loss: 0.0877 - val_accuracy: 0.9444 - lr: 3.0000e-04
epoch 22/1000
14/14 [==============================] - eta: 0s - loss: 0.0809 - accuracy: 0.9722
epoch 00022: reducelronplateau reducing learning rate to 9.000000427477062e-05.
14/14 [==============================] - 1s 95ms/step - loss: 0.0809 - accuracy: 0.9722 - val_loss: 0.0897 - val_accuracy: 0.9444 - lr: 3.0000e-04
epoch 23/1000
14/14 [==============================] - 1s 95ms/step - loss: 0.0677 - accuracy: 0.9792 - val_loss: 0.0834 - val_accuracy: 0.9537 - lr: 9.0000e-05
epoch 24/1000
14/14 [==============================] - 1s 93ms/step - loss: 0.0741 - accuracy: 0.9722 - val_loss: 0.0771 - val_accuracy: 0.9537 - lr: 9.0000e-05
epoch 25/1000
14/14 [==============================] - 1s 94ms/step - loss: 0.0672 - accuracy: 0.9815 - val_loss: 0.0733 - val_accuracy: 0.9537 - lr: 9.0000e-05
epoch 26/1000
14/14 [==============================] - eta: 0s - loss: 0.0595 - accuracy: 0.9838
epoch 00026: reducelronplateau reducing learning rate to 2.700000040931627e-05.
14/14 [==============================] - 1s 95ms/step - loss: 0.0595 - accuracy: 0.9838 - val_loss: 0.0694 - val_accuracy: 0.9537 - lr: 9.0000e-05
epoch 27/1000
14/14 [==============================] - 1s 94ms/step - loss: 0.0631 - accuracy: 0.9838 - val_loss: 0.0699 - val_accuracy: 0.9537 - lr: 2.7000e-05
epoch 28/1000
14/14 [==============================] - 1s 97ms/step - loss: 0.0591 - accuracy: 0.9861 - val_loss: 0.0705 - val_accuracy: 0.9537 - lr: 2.7000e-05
epoch 29/1000
14/14 [==============================] - eta: 0s - loss: 0.0635 - accuracy: 0.9838
epoch 00029: reducelronplateau reducing learning rate to 8.100000013655517e-06.
14/14 [==============================] - 1s 95ms/step - loss: 0.0635 - accuracy: 0.9838 - val_loss: 0.0697 - val_accuracy: 0.9444 - lr: 2.7000e-05
epoch 30/1000
14/14 [==============================] - 1s 95ms/step - loss: 0.0643 - accuracy: 0.9792 - val_loss: 0.0687 - val_accuracy: 0.9444 - lr: 8.1000e-06
epoch 31/1000
14/14 [==============================] - 1s 100ms/step - loss: 0.0768 - accuracy: 0.9745 - val_loss: 0.0665 - val_accuracy: 0.9537 - lr: 8.1000e-06
epoch 32/1000
14/14 [==============================] - eta: 0s - loss: 0.0645 - accuracy: 0.9861
epoch 00032: reducelronplateau reducing learning rate to 2.429999949526973e-06.
14/14 [==============================] - 1s 95ms/step - loss: 0.0645 - accuracy: 0.9861 - val_loss: 0.0656 - val_accuracy: 0.9537 - lr: 8.1000e-06
epoch 33/1000
14/14 [==============================] - 1s 97ms/step - loss: 0.0635 - accuracy: 0.9792 - val_loss: 0.0645 - val_accuracy: 0.9630 - lr: 2.4300e-06
epoch 34/1000
14/14 [==============================] - 1s 95ms/step - loss: 0.0606 - accuracy: 0.9838 - val_loss: 0.0636 - val_accuracy: 0.9630 - lr: 2.4300e-06
epoch 35/1000
14/14 [==============================] - 1s 95ms/step - loss: 0.0620 - accuracy: 0.9907 - val_loss: 0.0628 - val_accuracy: 0.9630 - lr: 2.4300e-06
epoch 36/10009/14 [==================>...........] - eta: 0s - loss: 0.0729 - accuracy: 0.9826warning:tensorflow:can save best model only with val_accuracy available, skipping.
14/14 [==============================] - eta: 0s - loss: 0.0682 - accuracy: 0.9861
epoch 00036: reducelronplateau reducing learning rate to 7.289999985005124e-07.
14/14 [==============================] - 1s 95ms/step - loss: 0.0682 - accuracy: 0.9861 - val_loss: 0.0622 - val_accuracy: 0.9630 - lr: 2.4300e-06
epoch 37/1000
14/14 [==============================] - 1s 96ms/step - loss: 0.0573 - accuracy: 0.9907 - val_loss: 0.0613 - val_accuracy: 0.9630 - lr: 7.2900e-07
epoch 38/1000
14/14 [==============================] - 1s 97ms/step - loss: 0.0575 - accuracy: 0.9931 - val_loss: 0.0607 - val_accuracy: 0.9722 - lr: 7.2900e-07
epoch 39/1000
14/14 [==============================] - 1s 94ms/step - loss: 0.0622 - accuracy: 0.9769 - val_loss: 0.0600 - val_accuracy: 0.9722 - lr: 7.2900e-07
epoch 40/1000
14/14 [==============================] - 1s 96ms/step - loss: 0.0660 - accuracy: 0.9838 - val_loss: 0.0594 - val_accuracy: 0.9722 - lr: 7.2900e-07
epoch 41/1000
14/14 [==============================] - eta: 0s - loss: 0.0614 - accuracy: 0.9884
epoch 00041: reducelronplateau reducing learning rate to 2.1870000637136398e-07.
14/14 [==============================] - 1s 95ms/step - loss: 0.0614 - accuracy: 0.9884 - val_loss: 0.0591 - val_accuracy: 0.9722 - lr: 7.2900e-07
epoch 42/1000
14/14 [==============================] - 1s 94ms/step - loss: 0.0605 - accuracy: 0.9792 - val_loss: 0.0583 - val_accuracy: 0.9722 - lr: 2.1870e-07
epoch 43/1000
14/14 [==============================] - 1s 99ms/step - loss: 0.0529 - accuracy: 0.9954 - val_loss: 0.0582 - val_accuracy: 0.9722 - lr: 2.1870e-07
epoch 44/1000
14/14 [==============================] - eta: 0s - loss: 0.0500 - accuracy: 0.9884
epoch 00044: reducelronplateau reducing learning rate to 6.561000276406048e-08.
14/14 [==============================] - 1s 95ms/step - loss: 0.0500 - accuracy: 0.9884 - val_loss: 0.0580 - val_accuracy: 0.9722 - lr: 2.1870e-07
epoch 45/1000
14/14 [==============================] - 1s 94ms/step - loss: 0.0613 - accuracy: 0.9861 - val_loss: 0.0581 - val_accuracy: 0.9722 - lr: 6.5610e-08
epoch 46/1000
14/14 [==============================] - 1s 94ms/step - loss: 0.0672 - accuracy: 0.9861 - val_loss: 0.0572 - val_accuracy: 0.9722 - lr: 6.5610e-08
epoch 47/1000
14/14 [==============================] - eta: 0s - loss: 0.0511 - accuracy: 0.9931
epoch 00047: reducelronplateau reducing learning rate to 1.9683000829218145e-08.
14/14 [==============================] - 1s 96ms/step - loss: 0.0511 - accuracy: 0.9931 - val_loss: 0.0574 - val_accuracy: 0.9722 - lr: 6.5610e-08
epoch 48/1000
14/14 [==============================] - 1s 99ms/step - loss: 0.0622 - accuracy: 0.9861 - val_loss: 0.0570 - val_accuracy: 0.9722 - lr: 1.9683e-08
plt
.plot
(history
.history
["accuracy"])
plt
.plot
(history
.history
["val_accuracy"])
plt
.title
("model accuracy")
plt
.ylabel
("accuracy")
plt
.xlabel
("epoch")
plt
.legend
(["train", "test"], loc
= "upper left")
plt
.show
()
plt
.plot
(history
.history
["loss"])
plt
.plot
(history
.history
["val_loss"])
plt
.title
("model loss")
plt
.ylabel
("loss")
plt
.xlabel
("epoch")
plt
.legend
(["train", "test"], loc
= "upper left")
plt
.show
()
def predict_class(img
):img
= img
.reshape
(1,128,128,3)predictions
= model
.predict
(img
)true_prediction
= [tf
.argmax
(pred
) for pred
in predictions
]true_prediction
= np
.array
(true_prediction
)return list(categories
.keys
())[list(categories
.values
()).index
(true_prediction
)]
y_pred
= model
.predict
(x_test
)
y_pred_
= [np
.argmax
(y
) for y
in y_pred
]
y_test_
= [np
.argmax
(y
) for y
in y_test
]
from sklearn
.metrics
import classification_report
print(classification_report
(y_test_
, y_pred_
))
precision recall f1-score support0 1.00 0.96 0.98 251 0.85 1.00 0.92 112 1.00 0.96 0.98 24accuracy 0.97 60macro avg 0.95 0.97 0.96 60
weighted avg 0.97 0.97 0.97 60
plt
.figure
(figsize
=(10,10))
random_inds
= np
.random
.choice
(x_test
.shape
[0],36)
for i
in range(36):plt
.subplot
(6,6,i
1)plt
.xticks
([])plt
.yticks
([])plt
.grid
(false)image_ind
= random_inds
[i
]plt
.imshow
(np
.squeeze
(x_test
[image_ind
]), cmap
=plt
.cm
.binary
)label
= predict_class
(x_test
[image_ind
])plt
.xlabel
(label
)
model
.save
("model.h5")
def predict(path
,model_str
,img_size
= (128,128)):new_model
= tf
.keras
.models
.load_model
(model_str
)img
= cv2
.imread
(path
)img
= cv2
.resize
(img
, img_size
) img
= np
.array
(img
) img
= img
.reshape
(1,128,128,3)predictions
= new_model
.predict
(img
)true_prediction
= [tf
.argmax
(pred
) for pred
in predictions
]true_prediction
= np
.array
(true_prediction
)return list(categories
.keys
())[list(categories
.values
()).index
(true_prediction
)]predict
("./data/a/051.jpg","model.h5")
'a'
predict
("./data/b/048.jpg","model.h5")
'b'
predict
("./data/c/050.jpg","model.h5")
'c'
代码资料下载: https://download.csdn.net/download/weixin_44510615/72775830
总结
以上是凯发k8官方网为你收集整理的深度学习图片分类cnn模板的全部内容,希望文章能够帮你解决所遇到的问题。
如果觉得凯发k8官方网网站内容还不错,欢迎将凯发k8官方网推荐给好友。