naiborhujosua

music

naiborhujosua

progress

naiborhujosua

year to date workouts

naiborhujosua

Progress running

naiborhujosua

# Make preds on a series of random images
import os
import random

plt.figure(figsize=(20, 14))
for i in range(3):
  # Choose a random image from a random class 
    class_name = random.choice(class_names)
    filename = random.choice(os.listdir(test_dir + "/" + class_name))
    filepath = test_dir + class_name + "/" + filename

naiborhujosua

# Plot a confusion matrix with all 3600 predictions, ground truth labels and 6 classes
make_confusion_matrix(y_true=y_labels,
                      y_pred=pred_classes,
                      classes=class_names_test,
                      figsize=(100, 100),
                      text_size=30,
                      norm=False,
                      savefig=True)

naiborhujosua

# Get accuracy score by comparing predicted classes to ground truth labels
from sklearn.metrics import accuracy_score
sklearn_accuracy = accuracy_score(y_labels, pred_classes)
sklearn_accuracy

naiborhujosua

# 1. Create base model with tf.keras.applications
base_model = tf.keras.applications.EfficientNetB0(include_top=False)

# 2. Freeze the base model (so the pre-learned patterns remain)
base_model.trainable = False

# 3. Create inputs into the base model
inputs = tf.keras.layers.Input(shape=(224, 224, 3), name="input_layer")

# 4. If using ResNet50V2, add this to speed up convergence, remove for EfficientNet

naiborhujosua

# Dataset split
train_dir = "/content/zw4p9kj6nt-2/data_splitting/Train/"
valid_dir ="/content/zw4p9kj6nt-2/data_splitting/Pred/"
test_dir = "/content/zw4p9kj6nt-2/data_splitting/Test/"
# Create data inputs
IMG_SIZE = (224, 224) # define image size
train_data = tf.keras.preprocessing.image_dataset_from_directory(directory=train_dir,
                                                                            image_size=IMG_SIZE,
                                                                            label_mode="categorical", # what type are the labels?
                                                                            batch_size=32) # batch_size is 32 by default, this is generally a good number

naiborhujosua

from pandas.plotting import parallel_coordinates

# Plot
plt.figure(figsize=(12,9), dpi= 80)
data.drop(['Date','Holiday', 'Functioning Day'],axis=1,inplace=True)
parallel_coordinates(data, 'Seasons', colormap='Dark2')

# Lighten borders
plt.gca().spines["top"].set_alpha(0)
plt.gca().spines["bottom"].set_alpha(.3)