!pip install -q kaggle
!mkdir -p ~/.kaggle
!cp kaggle.json ~/.kaggle/
!chmod 600 /root/.kaggle/kaggle.json
!kaggle datasets list
ref title size lastUpdated downloadCount voteCount usabilityRating ------------------------------------------------------------ ------------------------------------------------- ----- ------------------- ------------- --------- --------------- meirnizri/covid19-dataset COVID-19 Dataset 5MB 2022-11-13 15:47:17 15711 443 1.0 devrimtuner/list-of-moststreamed-songs-on-spotify Top 100 Spotify Songs👑🎤🎧🎼 3KB 2022-12-30 05:42:54 608 39 1.0 thedevastator/analyzing-credit-card-spending-habits-in-india Credit Card Spending Habits in India 319KB 2022-12-14 07:30:37 1878 63 1.0 die9origephit/fifa-world-cup-2022-complete-dataset Fifa World Cup 2022: Complete Dataset 7KB 2022-12-18 22:51:11 3647 127 1.0 michals22/coffee-dataset Coffee dataset 24KB 2022-12-15 20:02:12 3991 90 1.0 heemalichaudhari/netflix-movies-and-series Netflix Movies and Series 2MB 2022-12-22 13:34:22 1226 32 0.8235294 sejungjenn/spotify-best-songs-of-2022 Spotify: Winner Tracks Audio Features🎹 38KB 2022-12-28 08:06:49 315 22 1.0 thedevastator/unlock-profits-with-e-commerce-sales-data E-Commerce Sales Dataset 6MB 2022-12-03 09:27:17 3269 79 1.0 aklimarimi/qs-world-ranked-universities-20182022 QS World ranked Universities (2018-2022) 51KB 2022-12-28 03:53:39 650 32 1.0 rajeshrampure/black-friday-sale Black Friday Sale 5MB 2022-12-24 09:37:49 1029 32 1.0 devrimtuner/highestpaid-athletes HIGHEST-PAID ATHLETES⚽️🏀🏈⚾️🥎🎾 1KB 2022-12-29 01:29:51 312 27 1.0 heemalichaudhari/shopping Shopping 12KB 2022-12-26 14:25:07 515 28 0.9411765 milanvaddoriya/old-car-price-prediction Old car price prediction 105KB 2022-12-24 15:38:56 489 30 1.0 thedevastator/how-does-daily-yoga-impact-screen-time-habits How Does Daily Yoga Impact Screen Time Habits 742B 2022-12-14 04:10:56 738 24 1.0 thedevastator/uncovering-factors-that-affect-used-car-prices Used Cars 18MB 2022-12-06 13:36:08 1196 36 1.0 devrimtuner/list-of-mostfollowed-instagram-accounts (TOP 50)List of most-followed Instagram accounts👑 2KB 2022-12-30 07:52:00 366 27 1.0 thedevastator/jobs-dataset-from-glassdoor Salary Prediction 3MB 2022-11-16 13:52:31 8568 182 1.0 dansbecker/melbourne-housing-snapshot Melbourne Housing Snapshot 451KB 2018-06-05 12:52:24 103465 1191 0.7058824 mattop/best-selling-game-boy-video-games Best Selling Game Boy Video Games 2KB 2022-12-17 18:41:38 395 27 0.9705882 rajeshrampure/zomato-dataset Zomato Dataset 89MB 2022-12-23 07:38:07 628 28 1.0
!kaggle datasets download -d yaswanthgali/sport-celebrity-image-classification
Downloading sport-celebrity-image-classification.zip to /content 61% 9.00M/14.8M [00:00<00:00, 93.8MB/s] 100% 14.8M/14.8M [00:00<00:00, 128MB/s]
!unzip "/content/sport-celebrity-image-classification.zip" -d "/content/drive/MyDrive/Kaggle Datasets"
dir = "/content/drive/MyDrive/Kaggle Datasets/Sports-celebrity images"
from tensorflow.keras.preprocessing.image import ImageDataGenerator
# All images will be rescaled by 1./255.
datagen = ImageDataGenerator( rescale = 1.0/255. , validation_split=0.1,
rotation_range=20,
width_shift_range=0.2,
height_shift_range=0.2,
shear_range=60,
horizontal_flip=True)
train_generator = datagen.flow_from_directory(
directory=dir, target_size=(100, 100), color_mode='rgb', class_mode='categorical', batch_size=32,
shuffle=True, seed=2022, subset='training')
Found 399 images belonging to 4 classes.
validation_generator = datagen.flow_from_directory(
directory=dir, target_size=(100, 100), color_mode='rgb', class_mode='categorical',
batch_size=32, shuffle=True, seed=2022, subset='validation')
Found 42 images belonging to 4 classes.
import tensorflow as tf
tf.random.set_seed(2022)
model = tf.keras.models.Sequential([
tf.keras.layers.Conv2D(16, (3,3), activation='relu', input_shape=(100, 100, 3)),
tf.keras.layers.MaxPooling2D(2,1),
tf.keras.layers.Flatten(),
tf.keras.layers.Dense(1024, activation='relu'),
tf.keras.layers.BatchNormalization(),
tf.keras.layers.Dense(512, activation='relu'),
tf.keras.layers.BatchNormalization(),
tf.keras.layers.Dense(4, activation='softmax')
])
from tensorflow.keras.callbacks import EarlyStopping
monitor = EarlyStopping(monitor='val_loss', min_delta=1e-3, patience=5,verbose=3,restore_best_weights=True)
model.compile(optimizer='adam',loss='categorical_crossentropy', metrics = ['accuracy'])
history = model.fit(train_generator, validation_data=validation_generator, epochs=50,
verbose=2, callbacks=[monitor])
/usr/local/lib/python3.8/dist-packages/keras/preprocessing/image.py:1663: UserWarning: This ImageDataGenerator specifies `featurewise_center`, but it hasn't been fit on any training data. Fit it first by calling `.fit(numpy_data)`. warnings.warn('This ImageDataGenerator specifies ' /usr/local/lib/python3.8/dist-packages/keras/preprocessing/image.py:1671: UserWarning: This ImageDataGenerator specifies `featurewise_std_normalization`, but it hasn't been fit on any training data. Fit it first by calling `.fit(numpy_data)`. warnings.warn('This ImageDataGenerator specifies '
Epoch 1/50 13/13 - 28s - loss: 1.7844 - accuracy: 0.3860 - val_loss: 6.7128 - val_accuracy: 0.3810 - 28s/epoch - 2s/step Epoch 2/50 13/13 - 26s - loss: 1.1777 - accuracy: 0.5388 - val_loss: 3.2453 - val_accuracy: 0.3333 - 26s/epoch - 2s/step Epoch 3/50 13/13 - 26s - loss: 1.0512 - accuracy: 0.5514 - val_loss: 3.0801 - val_accuracy: 0.4524 - 26s/epoch - 2s/step Epoch 4/50 13/13 - 26s - loss: 1.0572 - accuracy: 0.5664 - val_loss: 2.9439 - val_accuracy: 0.4286 - 26s/epoch - 2s/step Epoch 5/50 13/13 - 26s - loss: 0.9758 - accuracy: 0.5940 - val_loss: 1.2449 - val_accuracy: 0.5952 - 26s/epoch - 2s/step Epoch 6/50 13/13 - 26s - loss: 0.8903 - accuracy: 0.6491 - val_loss: 1.0821 - val_accuracy: 0.5714 - 26s/epoch - 2s/step Epoch 7/50 13/13 - 25s - loss: 0.9012 - accuracy: 0.6366 - val_loss: 1.7198 - val_accuracy: 0.5238 - 25s/epoch - 2s/step Epoch 8/50 13/13 - 25s - loss: 0.8918 - accuracy: 0.6366 - val_loss: 1.7779 - val_accuracy: 0.4048 - 25s/epoch - 2s/step Epoch 9/50 13/13 - 33s - loss: 0.9311 - accuracy: 0.6541 - val_loss: 1.2837 - val_accuracy: 0.5476 - 33s/epoch - 3s/step Epoch 10/50 13/13 - 29s - loss: 0.8496 - accuracy: 0.6617 - val_loss: 1.2064 - val_accuracy: 0.5714 - 29s/epoch - 2s/step Epoch 11/50 13/13 - 27s - loss: 0.8027 - accuracy: 0.6767 - val_loss: 0.9762 - val_accuracy: 0.6429 - 27s/epoch - 2s/step Epoch 12/50 13/13 - 25s - loss: 0.8506 - accuracy: 0.6591 - val_loss: 1.1437 - val_accuracy: 0.4762 - 25s/epoch - 2s/step Epoch 13/50 13/13 - 26s - loss: 0.7915 - accuracy: 0.6892 - val_loss: 1.1869 - val_accuracy: 0.5476 - 26s/epoch - 2s/step Epoch 14/50 13/13 - 26s - loss: 0.8202 - accuracy: 0.6867 - val_loss: 1.6896 - val_accuracy: 0.4524 - 26s/epoch - 2s/step Epoch 15/50 13/13 - 25s - loss: 0.8408 - accuracy: 0.6617 - val_loss: 0.9959 - val_accuracy: 0.6190 - 25s/epoch - 2s/step Epoch 16/50 Restoring model weights from the end of the best epoch: 11. 13/13 - 26s - loss: 0.7455 - accuracy: 0.6842 - val_loss: 1.0675 - val_accuracy: 0.5476 - 26s/epoch - 2s/step Epoch 16: early stopping