search

nisargptl / vehicle-recognition

This enables users to gather information about any vehicle and they can see the details in a matter of seconds. Snap a quick photo in the app and it will provide the details about the vehicle make, model, year, user ratings and starting retail price. In addition, it will provide this same information for the 2 closest competitors so a user can do some quick comparison shopping.
https://viris.herokuapp.com/
0 stars 1 forks source link

increasing further accuracy #31

Closed khyatibhuva closed 3 years ago

khyatibhuva commented 3 years ago

As we had discussed in meaning, we were again able to improve accuracy. Just mention stats of the same and upload the updated code her for reference. You may also generate a pull request if needed.

201801196 commented 3 years ago

New validation accuracy is 90.20% for 100 classes.

import tensorflow as tf from tensorflow.keras import from tensorflow.keras.layers import from tensorflow.keras.models import Model from tensorflow.keras.preprocessing import image from tensorflow.keras.models import load_model from tensorflow.keras.preprocessing.image import ImageDataGenerator from tensorflow.keras.models import Sequential from keras.applications.mobilenet import MobileNet, preprocess_input import numpy as np from glob import glob import matplotlib.pyplot as plt import os

os.environ['TF_XLA_FLAGS'] = '--tf_xla_enable_xla_devices' tf.version assert tf.test.is_built_with_cuda() gpus = tf.config.list_physical_devices('GPU')

if gpus: try: tf.config.experimental.set_visible_devices(gpus[0], 'GPU') logical_gpus = tf.config.experimental.list_logical_devices('GPU') print(len(gpus), "Physical GPUs,", len(logical_gpus), "Logical GPU") except RuntimeError as e: print(e)

model=tf.keras.models.load_model('D:/car196/car100/New_model_100_classes_V2.h5')

for layer in model.layers: layer.trainable = False

for layer in model.layers[-30:]: layer.trainable = True

model.summary()

adam=tf.keras.optimizers.Adam( learning_rate=1e-5, beta_1=0.9, beta_2=0.999, epsilon=1e-07, decay = 0.3 )

model.compile( loss='categorical_crossentropy', optimizer="adam", metrics=['accuracy'] )

from tensorflow.keras.preprocessing.image import ImageDataGenerator

train_datagen = ImageDataGenerator(preprocessing_function=preprocess_input, rotation_range = 30, horizontal_flip = True)

test_datagen = ImageDataGenerator(preprocessing_function=preprocess_input) training_set = train_datagen.flow_from_directory('D:/car196/car100/training', target_size = (227, 227), batch_size = 96, class_mode = 'categorical', shuffle=True) test_set = test_datagen.flow_from_directory('D:/car196/car100/testing', target_size = (227, 227), batch_size = 64, class_mode = 'categorical', shuffle=True)

callback = [tf.keras.callbacks.EarlyStopping(monitor='loss', patience=4,mode = 'auto', restore_best_weights=False), tf.keras.callbacks.ReduceLROnPlateau(monitor="val_loss",factor=0.1, patience=2, mode="auto",min_delta=0.0001,min_lr=0), ] r = model.fit( training_set, validation_data=test_set, epochs=15, steps_per_epoch=len(training_set), validation_steps=len(test_set), callbacks = [callback] )

plt.plot(r.history['loss'], label='train loss') plt.plot(r.history['val_loss'], label='val loss') plt.legend() plt.show() plt.plot(r.history['accuracy'], label='train acc') plt.plot(r.history['val_accuracy'], label='val acc') plt.legend() plt.show()

model.save('D:/car196/car100/New_model_100_classes_V3.h5')

model.evaluate(test_set)

khyatibhuva commented 3 years ago

Great work. Noted.