In [2]:

    

with open('./food101/calories.txt') as f:
    calories = [float(y) for _, y in map(str.split, f)]


    

In [9]:

    

import matplotlib.pyplot as plt
import numpy as np
plt.hist(calories, bins=100)
plt.show()


    

In [11]:

    

np.mean(calories), np.std(calories)


    

    
Out[11]:





(237.05069306930693, 164.14514985917779)

In [12]:

    

calories = np.array(calories)


    

In [17]:

    

mse_of_mean = np.mean((calories - np.mean(calories)) ** 2)
print('mse of mean', mse_of_mean)
avg_percent = np.mean( np.abs(calories - np.mean(calories)) / calories)
print('avg_percent', avg_percent)


    

    




mse of mean 26943.6302223
avg_percent 0.893165039281

In [31]:

    

sampled = np.random.normal(np.mean(calories), np.std(calories), len(calories))
avg_percent = np.mean( np.abs(calories - sampled) / calories)
print('avg_percent', avg_percent)


    

    




avg_percent 1.25104638552

In [33]:

    

np.argmax(calories > 1000)


    

    
Out[33]:





31000

In [36]:

    

import scipy.misc as sm
im = sm.imread('./food101/31000.jpg')
plt.imshow(im)
plt.show()
print(calories[31000])


    

    













    




1388.0