In [1]:
from google.colab import drive
drive.mount('/content/drive')
Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount("/content/drive", force_remount=True).
In [2]:
cd/content/drive/MyDrive/AI_IBM
/content/drive/MyDrive/AI_IBM
In [3]:
!unzip Flowers-Dataset.zip
Archive: Flowers-Dataset.zip replace flowers/daisy/100080576_f52e8ee070_n.jpg? [y]es, [n]o, [A]ll, [N]one, [r]ename: N
In [4]:
from tensorflow.keras.preprocessing.image import ImageDataGenerator
In [5]:
train_datagen=ImageDataGenerator(rescale=1./255, zoom_range=0.2,horizontal_flip=True,vertical_flip=False)
In [6]:
test_datagen=ImageDataGenerator(rescale=1./255)
In [7]:
x_train=train_datagen.flow_from_directory(r"/content/drive/MyDrive/AI_IBM/flowers",target_size=(64,64),class_mode='categorical',batch_size=24)
Found 4317 images belonging to 5 classes.
In [8]:
x_test=test_datagen.flow_from_directory(r"/content/drive/MyDrive/AI_IBM/flowers",target_size=(64,64),class_mode='categorical',batch_size=24)
Found 4317 images belonging to 5 classes.
In [9]:
x_train.class_indices
Out[9]:
{'daisy': 0, 'dandelion': 1, 'rose': 2, 'sunflower': 3, 'tulip': 4}
In [10]:
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense,Convolution2D,MaxPooling2D,Flatten
In [11]:
model=Sequential()
In [12]:
model.add(Convolution2D(32,(3,3),input_shape=(64,64,3),activation='relu'))
In [13]:
model.add(MaxPooling2D(pool_size=(2,2)))
In [14]:
model.add(Flatten())
In [15]:
model.summary()
Model: "sequential"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
conv2d (Conv2D) (None, 62, 62, 32) 896
max_pooling2d (MaxPooling2D (None, 31, 31, 32) 0
)
flatten (Flatten) (None, 30752) 0
=================================================================
Total params: 896
Trainable params: 896
Non-trainable params: 0
_________________________________________________________________
In [16]:
model.add(Dense(300,activation='relu'))
model.add(Dense(150,activation='relu'))
In [17]:
model.add(Dense(5,activation='softmax'))
In [18]:
model.compile(loss='categorical_crossentropy',optimizer='adam',metrics=['accuracy'])
In [19]:
len(x_train)
Out[19]:
180
In [20]:
model.fit_generator(x_train,steps_per_epoch=len(x_train), validation_data=x_test, validation_steps=len(x_test), epochs= 30)
/usr/local/lib/python3.7/dist-packages/ipykernel_launcher.py:1: UserWarning: `Model.fit_generator` is deprecated and will be removed in a future version. Please use `Model.fit`, which supports generators. """Entry point for launching an IPython kernel.
Epoch 1/30 180/180 [==============================] - 393s 2s/step - loss: 1.3213 - accuracy: 0.4714 - val_loss: 1.1275 - val_accuracy: 0.5532 Epoch 2/30 180/180 [==============================] - 74s 409ms/step - loss: 1.0600 - accuracy: 0.5854 - val_loss: 0.9406 - val_accuracy: 0.6301 Epoch 3/30 180/180 [==============================] - 73s 405ms/step - loss: 0.9678 - accuracy: 0.6247 - val_loss: 0.9603 - val_accuracy: 0.6203 Epoch 4/30 180/180 [==============================] - 77s 429ms/step - loss: 0.8884 - accuracy: 0.6546 - val_loss: 0.8187 - val_accuracy: 0.6938 Epoch 5/30 180/180 [==============================] - 76s 422ms/step - loss: 0.8358 - accuracy: 0.6787 - val_loss: 0.7393 - val_accuracy: 0.7225 Epoch 6/30 180/180 [==============================] - 75s 418ms/step - loss: 0.7924 - accuracy: 0.6965 - val_loss: 0.8389 - val_accuracy: 0.6928 Epoch 7/30 180/180 [==============================] - 73s 405ms/step - loss: 0.7521 - accuracy: 0.7158 - val_loss: 0.8503 - val_accuracy: 0.6789 Epoch 8/30 180/180 [==============================] - 74s 411ms/step - loss: 0.7048 - accuracy: 0.7313 - val_loss: 0.6492 - val_accuracy: 0.7521 Epoch 9/30 180/180 [==============================] - 72s 400ms/step - loss: 0.6502 - accuracy: 0.7521 - val_loss: 0.6458 - val_accuracy: 0.7438 Epoch 10/30 180/180 [==============================] - 74s 409ms/step - loss: 0.6182 - accuracy: 0.7684 - val_loss: 0.5721 - val_accuracy: 0.7818 Epoch 11/30 180/180 [==============================] - 72s 402ms/step - loss: 0.5662 - accuracy: 0.7931 - val_loss: 0.5968 - val_accuracy: 0.7725 Epoch 12/30 180/180 [==============================] - 72s 401ms/step - loss: 0.5600 - accuracy: 0.7908 - val_loss: 0.6907 - val_accuracy: 0.7612 Epoch 13/30 180/180 [==============================] - 72s 399ms/step - loss: 0.5064 - accuracy: 0.8138 - val_loss: 0.5185 - val_accuracy: 0.8117 Epoch 14/30 180/180 [==============================] - 71s 394ms/step - loss: 0.4830 - accuracy: 0.8249 - val_loss: 0.3613 - val_accuracy: 0.8673 Epoch 15/30 180/180 [==============================] - 71s 397ms/step - loss: 0.4650 - accuracy: 0.8196 - val_loss: 0.3396 - val_accuracy: 0.8768 Epoch 16/30 180/180 [==============================] - 71s 393ms/step - loss: 0.4117 - accuracy: 0.8559 - val_loss: 0.3472 - val_accuracy: 0.8738 Epoch 17/30 180/180 [==============================] - 71s 397ms/step - loss: 0.3892 - accuracy: 0.8631 - val_loss: 0.3314 - val_accuracy: 0.8826 Epoch 18/30 180/180 [==============================] - 70s 389ms/step - loss: 0.3441 - accuracy: 0.8726 - val_loss: 0.4008 - val_accuracy: 0.8589 Epoch 19/30 180/180 [==============================] - 73s 404ms/step - loss: 0.3467 - accuracy: 0.8719 - val_loss: 0.2484 - val_accuracy: 0.9060 Epoch 20/30 180/180 [==============================] - 72s 398ms/step - loss: 0.3327 - accuracy: 0.8758 - val_loss: 0.2234 - val_accuracy: 0.9210 Epoch 21/30 180/180 [==============================] - 73s 403ms/step - loss: 0.2807 - accuracy: 0.9009 - val_loss: 0.2830 - val_accuracy: 0.9036 Epoch 22/30 180/180 [==============================] - 70s 392ms/step - loss: 0.2751 - accuracy: 0.9013 - val_loss: 0.2392 - val_accuracy: 0.9141 Epoch 23/30 180/180 [==============================] - 73s 404ms/step - loss: 0.2549 - accuracy: 0.9097 - val_loss: 0.2221 - val_accuracy: 0.9189 Epoch 24/30 180/180 [==============================] - 72s 399ms/step - loss: 0.2412 - accuracy: 0.9243 - val_loss: 0.2029 - val_accuracy: 0.9291 Epoch 25/30 180/180 [==============================] - 72s 402ms/step - loss: 0.2360 - accuracy: 0.9199 - val_loss: 0.1965 - val_accuracy: 0.9307 Epoch 26/30 180/180 [==============================] - 72s 401ms/step - loss: 0.2199 - accuracy: 0.9201 - val_loss: 0.1919 - val_accuracy: 0.9331 Epoch 27/30 180/180 [==============================] - 72s 400ms/step - loss: 0.2008 - accuracy: 0.9363 - val_loss: 0.1218 - val_accuracy: 0.9560 Epoch 28/30 180/180 [==============================] - 73s 406ms/step - loss: 0.1889 - accuracy: 0.9310 - val_loss: 0.2838 - val_accuracy: 0.9108 Epoch 29/30 180/180 [==============================] - 70s 389ms/step - loss: 0.2046 - accuracy: 0.9275 - val_loss: 0.2116 - val_accuracy: 0.9307 Epoch 30/30 180/180 [==============================] - 70s 392ms/step - loss: 0.1886 - accuracy: 0.9372 - val_loss: 0.2091 - val_accuracy: 0.9280
Out[20]:
In [21]:
model.save('Flowers_classification_model1.h5')
In [22]:
ls
flowers/ Flowers_classification_model1.h5 Flowers-Dataset.zip video.mp4
In [23]:
import numpy as np
from tensorflow.keras.models import load_model
from tensorflow.keras.preprocessing import image
In [24]:
# Load the model
model=load_model('Flowers_classification_model1.h5')
In [38]:
img=image.load_img(r"/content/s3.jpg",target_size=(64,64))
x=image.img_to_array(img)
x=np.expand_dims(x,axis=0)
y=np.argmax(model.predict(x),axis=1)
# x_train.class_indices
index=['daisy','dandelion','rose','sunflower','tulip']
index[y[0]]
Out[38]:
'sunflower'
Copyright © Code Fetcher 2022