Increasing loss funcion in elementary example - python

I want to predict ideal linear data (identical function)
data = np.asarray(range(100),dtype=np.float32)
I use to this linear function
model = Sequential([
Dense(1, input_shape=(1,))
])
model.compile(optimizer='sgd', loss='mse')
model.fit(data, data, epochs=10, batch_size=100)
but my loss function is increasing. What is wrong with this simple code?
Epoch 1/10
100/100 [==============================] - 1s 7ms/step - loss: 3559.4075
Epoch 2/10
100/100 [==============================] - 0s 20us/step - loss: 14893056.0000
Epoch 3/10
100/100 [==============================] - 0s 170us/step - loss: 62314639360.0000
Epoch 4/10
100/100 [==============================] - 0s 30us/step - loss: 260733187129344.0000
Epoch 5/10
100/100 [==============================] - 0s 70us/step - loss: 1090944439330799616.0000
Epoch 6/10
100/100 [==============================] - 0s 20us/step - loss: 4564665060617919397888.0000
Epoch 7/10
100/100 [==============================] - 0s 30us/step - loss: 19099198494067630815576064.0000
Epoch 8/10
100/100 [==============================] - 0s 30us/step - loss: 79913699011849558249925771264.0000
Epoch 9/10
100/100 [==============================] - 0s 50us/step - loss: 334370041805433555342669660553216.0000
Epoch 10/10
100/100 [==============================] - 0s 20us/step - loss: 1399051141583436919510296595359858688.0000

You need to standardize input features. And you can learn How and why do normalization and feature scaling work?. Let me just use (x-mean(x))/std(x)as an example here.
import numpy as np
from keras.layers import Dense
from keras.models import Sequential
data = np.asarray(range(100),dtype=np.float32)
model = Sequential([
Dense(1, input_shape=(1,))
])
model.compile(optimizer='sgd', loss='mse')
model.fit((data-np.mean(data))/np.std(data), data, epochs=200, batch_size=100)
Epoch 1/200
100/100 [==============================] - 3s 26ms/step - loss: 3284.6235
Epoch 2/200
100/100 [==============================] - 0s 25us/step - loss: 3154.5522
Epoch 3/200
100/100 [==============================] - 0s 22us/step - loss: 3029.6318
...
100/100 [==============================] - 0s 27us/step - loss: 1.1016
Epoch 200/200
100/100 [==============================] - 0s 28us/step - loss: 1.0579

Related

LSTM has high accuracy while training, however, when I test it on the training set it seems to do very poorly

So I am trying to create an LSTM that can predict the next time step of a double pendulum. The data that I am trying to train with is a (2001, 4) numpy array. (i.e. the first 5 rows will look like:
array([[ 1.04719755, 0. , 1.04719755, 0. ],
[ 1.03659984, -0.42301933, 1.04717544, -0.00178865],
[ 1.00508218, -0.83475539, 1.04682248, -0.01551541],
[ 0.95354768, -1.22094052, 1.04514269, -0.05838011],
[ 0.88372305, -1.56345555, 1.04009056, -0.15443162]])
where each row is a unique representation of the state of the double pendulum.)
So I wanted to created an LSTM that could learn to predict the next state given the current one.
Here was my code for it so far (full_sol is the (2001, 4) matrix:
import numpy as np
from tensorflow import keras
import tensorflow as tf
# full_sol = np.random.rand(2001, 4)
full_sol = full_sol.reshape((full_sol.shape[0], 1, full_sol.shape[1]))
model = keras.Sequential()
model.add(keras.layers.LSTM(100, input_shape=(None, 4), return_sequences=True, dropout=0.2))
model.add(keras.layers.TimeDistributed(keras.layers.Dense(4, activation=tf.keras.layers.LeakyReLU(
alpha=0.3))))
model.compile(loss="mean_squared_error", optimizer="adam", metrics="accuracy")
history = model.fit(full_sol[:-1,:,:], full_sol[1:,:,:], epochs=20)
Then when I train, I get the following results:
Epoch 1/20
63/63 [==============================] - 3s 4ms/step - loss: 1.7181 - accuracy: 0.4200
Epoch 2/20
63/63 [==============================] - 0s 4ms/step - loss: 1.0481 - accuracy: 0.5155
Epoch 3/20
63/63 [==============================] - 0s 5ms/step - loss: 0.7584 - accuracy: 0.5715
Epoch 4/20
63/63 [==============================] - 0s 5ms/step - loss: 0.5134 - accuracy: 0.6420
Epoch 5/20
63/63 [==============================] - 0s 5ms/step - loss: 0.3944 - accuracy: 0.7260
Epoch 6/20
63/63 [==============================] - 0s 5ms/step - loss: 0.3378 - accuracy: 0.7605
Epoch 7/20
63/63 [==============================] - 0s 5ms/step - loss: 0.3549 - accuracy: 0.7825
Epoch 8/20
63/63 [==============================] - 0s 4ms/step - loss: 0.3528 - accuracy: 0.7995
Epoch 9/20
63/63 [==============================] - 0s 5ms/step - loss: 0.3285 - accuracy: 0.8020
Epoch 10/20
63/63 [==============================] - 0s 5ms/step - loss: 0.2874 - accuracy: 0.8030
Epoch 11/20
63/63 [==============================] - 0s 4ms/step - loss: 0.3072 - accuracy: 0.8135
Epoch 12/20
63/63 [==============================] - 0s 4ms/step - loss: 0.3075 - accuracy: 0.8035
Epoch 13/20
63/63 [==============================] - 0s 4ms/step - loss: 0.2942 - accuracy: 0.8030
Epoch 14/20
63/63 [==============================] - 0s 4ms/step - loss: 0.2637 - accuracy: 0.8170
Epoch 15/20
63/63 [==============================] - 0s 4ms/step - loss: 0.2675 - accuracy: 0.8150
Epoch 16/20
63/63 [==============================] - 0s 4ms/step - loss: 0.2644 - accuracy: 0.8085
Epoch 17/20
63/63 [==============================] - 0s 5ms/step - loss: 0.2479 - accuracy: 0.8200
Epoch 18/20
63/63 [==============================] - 0s 4ms/step - loss: 0.2475 - accuracy: 0.8215
Epoch 19/20
63/63 [==============================] - 0s 4ms/step - loss: 0.2243 - accuracy: 0.8340
Epoch 20/20
63/63 [==============================] - 0s 5ms/step - loss: 0.2430 - accuracy: 0.8240
So, quite high accuracy. But when I test it on the training set, the predictions aren't very good.
E.g. when I predict the first value:
model.predict(tf.expand_dims(full_sol[0], axis = 0))
I get array([[[ 1.0172144 , -0.3535697 , 1.1287913 , -0.23707283]]],dtype=float32)
Instead of array([[ 1.03659984, -0.42301933, 1.04717544, -0.00178865]]).
Where have I gone wrong?
I don't think you are doing anything wrong. What you are getting is still fairly close to the actual value. You can either change your choice of metric so it accurately represents the degree of error in your predictions, or you could try to increase the accuracy further.

NMT LSTM gives an incorrect response, and a big loss

I am writing a neural network for translating texts from Russian to English, but I ran into the problem that my neural network gives a big loss, as well as a very far from the correct answer.
Below is the LSTM that I build using Keras:
def make_model(in_vocab, out_vocab, in_timesteps, out_timesteps, n):
model = Sequential()
model.add(Embedding(in_vocab, n, input_length=in_timesteps, mask_zero=True))
model.add(LSTM(n))
model.add(Dropout(0.3))
model.add(RepeatVector(out_timesteps))
model.add(LSTM(n, return_sequences=True))
model.add(Dropout(0.3))
model.add(Dense(out_vocab, activation='softmax'))
model.compile(optimizer=optimizers.RMSprop(lr=0.001), loss='sparse_categorical_crossentropy')
return model
And also the learning process is presented:
Epoch 1/10
3/3 [==============================] - 5s 1s/step - loss: 8.3635 - accuracy: 0.0197 - val_loss: 8.0575 - val_accuracy: 0.0563
Epoch 2/10
3/3 [==============================] - 2s 806ms/step - loss: 7.9505 - accuracy: 0.0334 - val_loss: 8.2927 - val_accuracy: 0.0743
Epoch 3/10
3/3 [==============================] - 2s 812ms/step - loss: 7.7977 - accuracy: 0.0349 - val_loss: 8.2959 - val_accuracy: 0.0571
Epoch 4/10
3/3 [==============================] - 3s 825ms/step - loss: 7.6700 - accuracy: 0.0389 - val_loss: 8.5628 - val_accuracy: 0.0751
Epoch 5/10
3/3 [==============================] - 3s 829ms/step - loss: 7.5595 - accuracy: 0.0411 - val_loss: 8.5854 - val_accuracy: 0.0743
Epoch 6/10
3/3 [==============================] - 3s 807ms/step - loss: 7.4604 - accuracy: 0.0406 - val_loss: 8.7633 - val_accuracy: 0.0743
Epoch 7/10
3/3 [==============================] - 2s 815ms/step - loss: 7.3475 - accuracy: 0.0436 - val_loss: 8.9103 - val_accuracy: 0.0743
Epoch 8/10
3/3 [==============================] - 3s 825ms/step - loss: 7.2548 - accuracy: 0.0455 - val_loss: 9.0493 - val_accuracy: 0.0721
Epoch 9/10
3/3 [==============================] - 2s 814ms/step - loss: 7.1751 - accuracy: 0.0449 - val_loss: 9.0740 - val_accuracy: 0.0788
Epoch 10/10
3/3 [==============================] - 3s 831ms/step - loss: 7.1132 - accuracy: 0.0479 - val_loss: 9.2443 - val_accuracy: 0.0773
And the parameters that I transmit for training:
model = make_model(# the size of tokenized words
russian_vocab_size,
english_vocab_size,
# maximum sentence lengths
max_russian_sequence_length,
max_english_sequence_length,
512)
model.fit(preproc_russian_sentences, # all tokenized Russian offers that are transmitted in the format shape (X, Y)
preproc_english_sentences, # all tokenized English offers that are transmitted in the format shape (X, Y, 1)
epochs=10,
batch_size=1024,
validation_split=0.2,
callbacks=None,
verbose=1)
Thank you in advance.

Keras model does not generalise

I am trying to build a deep learning model on Keras for a test and I am not very good at this. I have a scaled dataset with 128 features and these correspond to 6 different classes.
I have already tried adding/deleting layers or using regularisation like dropout/l1/l2, My model learns and accuracy goes up so high. But accuracy on test set is around 15%.
from tensorflow.keras.layers import Dense, Dropout
model = Sequential()
model.add(Dense(128, activation='tanh', input_shape=(128,)))
model.add(Dropout(0.5))
model.add(Dense(60, activation='tanh'))
model.add(Dropout(0.5))
model.add(Dense(20, activation='tanh'))
model.add(Dropout(0.5))
model.add(Dense(6, activation='sigmoid'))
model.compile(loss='categorical_crossentropy', optimizer='Nadam', metrics=['accuracy'])
model.fit(train_X, train_y, epochs=20, batch_size=32, verbose=1)
6955/6955 [==============================] - 1s 109us/sample - loss: 1.5805 - acc: 0.3865
Epoch 2/20
6955/6955 [==============================] - 0s 71us/sample - loss: 1.1512 - acc: 0.6505
Epoch 3/20
6955/6955 [==============================] - 0s 71us/sample - loss: 0.9191 - acc: 0.7307
Epoch 4/20
6955/6955 [==============================] - 0s 67us/sample - loss: 0.7819 - acc: 0.7639
Epoch 5/20
6955/6955 [==============================] - 0s 66us/sample - loss: 0.6939 - acc: 0.7882
Epoch 6/20
6955/6955 [==============================] - 0s 69us/sample - loss: 0.6284 - acc: 0.8099
Epoch 7/20
6955/6955 [==============================] - 0s 70us/sample - loss: 0.5822 - acc: 0.8240
Epoch 8/20
6955/6955 [==============================] - 1s 73us/sample - loss: 0.5305 - acc: 0.8367
Epoch 9/20
6955/6955 [==============================] - 1s 75us/sample - loss: 0.5130 - acc: 0.8441
Epoch 10/20
6955/6955 [==============================] - 1s 75us/sample - loss: 0.4703 - acc: 0.8591
Epoch 11/20
6955/6955 [==============================] - 1s 73us/sample - loss: 0.4679 - acc: 0.8650
Epoch 12/20
6955/6955 [==============================] - 1s 77us/sample - loss: 0.4399 - acc: 0.8705
Epoch 13/20
6955/6955 [==============================] - 1s 80us/sample - loss: 0.4055 - acc: 0.8904
Epoch 14/20
6955/6955 [==============================] - 1s 77us/sample - loss: 0.3965 - acc: 0.8874
Epoch 15/20
6955/6955 [==============================] - 1s 77us/sample - loss: 0.3964 - acc: 0.8877
Epoch 16/20
6955/6955 [==============================] - 1s 77us/sample - loss: 0.3564 - acc: 0.9048
Epoch 17/20
6955/6955 [==============================] - 1s 80us/sample - loss: 0.3517 - acc: 0.9087
Epoch 18/20
6955/6955 [==============================] - 1s 78us/sample - loss: 0.3254 - acc: 0.9133
Epoch 19/20
6955/6955 [==============================] - 1s 78us/sample - loss: 0.3367 - acc: 0.9116
Epoch 20/20
6955/6955 [==============================] - 1s 76us/sample - loss: 0.3165 - acc: 0.9192
The result I am recieving 39% With other models like GBM or XGB I can reach upto 85%
What am I doing wrong? Any suggestions?

Neural net fails on toy dataset

I have created the following toy dataset:
I am trying to predict the class with a neural net in keras:
model = Sequential()
model.add(Dense(units=2, activation='sigmoid', input_shape= (nr_feats,)))
model.add(Dense(units=nr_classes, activation='softmax'))
model.compile(loss='categorical_crossentropy',
optimizer='adam',
metrics=['accuracy'])
With nr_feats and nr_classes set to 2.
The neural net can only predict with 50 percent accuracy returning either all 1's or all 2's. Using Logistic Regression results in 100 percent accuracy.
I can not find what is going wrong here.
I have uploaded a notebook to github if you quickly want to try something.
EDIT 1
I drastically increased the number of epochs and accuracy finally starts to improve from 0.5 at epoch 72 and converges to 1.0 at epoch 98.
This still seems extremely slow for such a simple dataset.
I am aware it is better to use a single output neuron with sigmoid activation but it's more that I want to understand why it does not work with two output neurons and softmax activation.
I pre-process my dataframe as follows:
from sklearn.preprocessing import LabelEncoder
x_train = df_train.iloc[:,0:-1].values
y_train = df_train.iloc[:, -1]
nr_feats = x_train.shape[1]
nr_classes = y_train.nunique()
label_enc = LabelEncoder()
label_enc.fit(y_train)
y_train = keras.utils.to_categorical(label_enc.transform(y_train), nr_classes)
Training and evaluation:
model.fit(x_train, y_train, epochs=500, batch_size=32, verbose=True)
accuracy_score(model.predict_classes(x_train), df_train.iloc[:, -1].values)
EDIT 2
After changing the output layer to a single neuron with sigmoid activation and using binary_crossentropy loss as modesitt suggested, accuracy still remains at 0.5 for 200 epochs and converges to 1.0 100 epochs later.
Note: Read the "Update" section at the end of my answer if you want the true reason. In this scenario, the other two reasons I have mentioned are only valid when the learning rate is set to a low value (less than 1e-3).
I put together some code. It is very similar to yours but I just cleaned it a little bit and made it simpler for myself. As you can see, I use a dense layer with one unit with a sigmoid activation function for the last layer and just change the optimizer from adam to rmsprop (it is not important that much, you can use adam if you like):
import numpy as np
import random
# generate random data with two features
n_samples = 200
n_feats = 2
cls0 = np.random.uniform(low=0.2, high=0.4, size=(n_samples,n_feats))
cls1 = np.random.uniform(low=0.5, high=0.7, size=(n_samples,n_feats))
x_train = np.concatenate((cls0, cls1))
y_train = np.concatenate((np.zeros((n_samples,)), np.ones((n_samples,))))
# shuffle data because all negatives (i.e. class "0") are first
# and then all positives (i.e. class "1")
indices = np.arange(x_train.shape[0])
np.random.shuffle(indices)
x_train = x_train[indices]
y_train = y_train[indices]
from keras.models import Sequential
from keras.layers import Dense
model = Sequential()
model.add(Dense(2, activation='sigmoid', input_shape=(n_feats,)))
model.add(Dense(1, activation='sigmoid'))
model.compile(loss='binary_crossentropy',
optimizer='rmsprop',
metrics=['accuracy'])
model.summary()
model.fit(x_train, y_train, epochs=5, batch_size=32, verbose=True)
Here is the output:
Layer (type) Output Shape Param #
=================================================================
dense_25 (Dense) (None, 2) 6
_________________________________________________________________
dense_26 (Dense) (None, 1) 3
=================================================================
Total params: 9
Trainable params: 9
Non-trainable params: 0
_________________________________________________________________
Epoch 1/5
400/400 [==============================] - 0s 966us/step - loss: 0.7013 - acc: 0.5000
Epoch 2/5
400/400 [==============================] - 0s 143us/step - loss: 0.6998 - acc: 0.5000
Epoch 3/5
400/400 [==============================] - 0s 137us/step - loss: 0.6986 - acc: 0.5000
Epoch 4/5
400/400 [==============================] - 0s 149us/step - loss: 0.6975 - acc: 0.5000
Epoch 5/5
400/400 [==============================] - 0s 132us/step - loss: 0.6966 - acc: 0.5000
As you can see the accuracy never increases from 50%. What if you increase the number of epochs to say 50:
Layer (type) Output Shape Param #
=================================================================
dense_35 (Dense) (None, 2) 6
_________________________________________________________________
dense_36 (Dense) (None, 1) 3
=================================================================
Total params: 9
Trainable params: 9
Non-trainable params: 0
_________________________________________________________________
Epoch 1/50
400/400 [==============================] - 0s 1ms/step - loss: 0.6925 - acc: 0.5000
Epoch 2/50
400/400 [==============================] - 0s 136us/step - loss: 0.6902 - acc: 0.5000
Epoch 3/50
400/400 [==============================] - 0s 133us/step - loss: 0.6884 - acc: 0.5000
Epoch 4/50
400/400 [==============================] - 0s 160us/step - loss: 0.6866 - acc: 0.5000
Epoch 5/50
400/400 [==============================] - 0s 140us/step - loss: 0.6848 - acc: 0.5000
Epoch 6/50
400/400 [==============================] - 0s 168us/step - loss: 0.6832 - acc: 0.5000
Epoch 7/50
400/400 [==============================] - 0s 154us/step - loss: 0.6817 - acc: 0.5000
Epoch 8/50
400/400 [==============================] - 0s 146us/step - loss: 0.6802 - acc: 0.5000
Epoch 9/50
400/400 [==============================] - 0s 161us/step - loss: 0.6789 - acc: 0.5000
Epoch 10/50
400/400 [==============================] - 0s 140us/step - loss: 0.6778 - acc: 0.5000
Epoch 11/50
400/400 [==============================] - 0s 177us/step - loss: 0.6766 - acc: 0.5000
Epoch 12/50
400/400 [==============================] - 0s 180us/step - loss: 0.6755 - acc: 0.5000
Epoch 13/50
400/400 [==============================] - 0s 165us/step - loss: 0.6746 - acc: 0.5000
Epoch 14/50
400/400 [==============================] - 0s 128us/step - loss: 0.6736 - acc: 0.5000
Epoch 15/50
400/400 [==============================] - 0s 125us/step - loss: 0.6728 - acc: 0.5000
Epoch 16/50
400/400 [==============================] - 0s 165us/step - loss: 0.6718 - acc: 0.5000
Epoch 17/50
400/400 [==============================] - 0s 161us/step - loss: 0.6710 - acc: 0.5000
Epoch 18/50
400/400 [==============================] - 0s 170us/step - loss: 0.6702 - acc: 0.5000
Epoch 19/50
400/400 [==============================] - 0s 122us/step - loss: 0.6694 - acc: 0.5000
Epoch 20/50
400/400 [==============================] - 0s 110us/step - loss: 0.6686 - acc: 0.5000
Epoch 21/50
400/400 [==============================] - 0s 142us/step - loss: 0.6676 - acc: 0.5000
Epoch 22/50
400/400 [==============================] - 0s 142us/step - loss: 0.6667 - acc: 0.5000
Epoch 23/50
400/400 [==============================] - 0s 149us/step - loss: 0.6659 - acc: 0.5000
Epoch 24/50
400/400 [==============================] - 0s 125us/step - loss: 0.6651 - acc: 0.5000
Epoch 25/50
400/400 [==============================] - 0s 134us/step - loss: 0.6643 - acc: 0.5000
Epoch 26/50
400/400 [==============================] - 0s 143us/step - loss: 0.6634 - acc: 0.5000
Epoch 27/50
400/400 [==============================] - 0s 137us/step - loss: 0.6625 - acc: 0.5000
Epoch 28/50
400/400 [==============================] - 0s 131us/step - loss: 0.6616 - acc: 0.5025
Epoch 29/50
400/400 [==============================] - 0s 119us/step - loss: 0.6608 - acc: 0.5100
Epoch 30/50
400/400 [==============================] - 0s 143us/step - loss: 0.6601 - acc: 0.5025
Epoch 31/50
400/400 [==============================] - 0s 148us/step - loss: 0.6593 - acc: 0.5350
Epoch 32/50
400/400 [==============================] - 0s 161us/step - loss: 0.6584 - acc: 0.5325
Epoch 33/50
400/400 [==============================] - 0s 152us/step - loss: 0.6576 - acc: 0.5700
Epoch 34/50
400/400 [==============================] - 0s 128us/step - loss: 0.6568 - acc: 0.5850
Epoch 35/50
400/400 [==============================] - 0s 155us/step - loss: 0.6560 - acc: 0.5975
Epoch 36/50
400/400 [==============================] - 0s 136us/step - loss: 0.6552 - acc: 0.6425
Epoch 37/50
400/400 [==============================] - 0s 140us/step - loss: 0.6544 - acc: 0.6150
Epoch 38/50
400/400 [==============================] - 0s 120us/step - loss: 0.6538 - acc: 0.6375
Epoch 39/50
400/400 [==============================] - 0s 140us/step - loss: 0.6531 - acc: 0.6725
Epoch 40/50
400/400 [==============================] - 0s 135us/step - loss: 0.6523 - acc: 0.6750
Epoch 41/50
400/400 [==============================] - 0s 136us/step - loss: 0.6515 - acc: 0.7300
Epoch 42/50
400/400 [==============================] - 0s 126us/step - loss: 0.6505 - acc: 0.7450
Epoch 43/50
400/400 [==============================] - 0s 141us/step - loss: 0.6496 - acc: 0.7425
Epoch 44/50
400/400 [==============================] - 0s 162us/step - loss: 0.6489 - acc: 0.7675
Epoch 45/50
400/400 [==============================] - 0s 161us/step - loss: 0.6480 - acc: 0.7775
Epoch 46/50
400/400 [==============================] - 0s 126us/step - loss: 0.6473 - acc: 0.7575
Epoch 47/50
400/400 [==============================] - 0s 124us/step - loss: 0.6464 - acc: 0.7625
Epoch 48/50
400/400 [==============================] - 0s 130us/step - loss: 0.6455 - acc: 0.7950
Epoch 49/50
400/400 [==============================] - 0s 191us/step - loss: 0.6445 - acc: 0.8100
Epoch 50/50
400/400 [==============================] - 0s 163us/step - loss: 0.6435 - acc: 0.8625
The accuracy starts to increase (Note that if you train this model multiple times, each time it may take different number of epochs to reach an acceptable accuracy, anything from 10 to 100 epochs).
Also, in my experiments I noticed that increasing the number of units in the first dense layer, for example to 5 or 10 units, causes the model to be trained faster (i.e. quickly converge).
Why so many epochs needed?
I think it is because of these two reasons (combined):
1) Despite the fact that the two classes are easily separable, your data is made up of random samples, and
2) The number of data points compared to the size of neural net (i.e. number of trainable parameters, which is 9 in example code above) is relatively large.
Therefore, it takes more epochs for the model to learn the weights. It is as though the model is very restricted and needs more and more experience to correctly find the appropriate weights. As an evidence, just try to increase the number of units in the first dense layer. You are almost guaranteed to reach an accuracy of +90% with less than 10 epochs each time you attempt to train this model. Here you increase the capacity and therefore the model converges (i.e. trains) much faster (it should be noted that it starts to overfit if the capacity is too high or you train the model for too many epochs. You should have a validation scheme to monitor this issue).
Side note:
Don't set the high argument to a number less than the low argument in numpy.random.uniform since, according to the documentation, the results will be "officially undefined" in this case.
Update:
One more important thing here (maybe the most important thing in this scenario) is the learning rate of the optimizer. If the learning rate is too low, the model converges slowly. Try increasing the learning rate, and you can see you reach an accuracy of 100% with less than 5 epochs:
from keras import optimizers
model.compile(loss='binary_crossentropy',
optimizer=optimizers.RMSprop(lr=1e-1),
metrics=['accuracy'])
# or you may use adam
model.compile(loss='binary_crossentropy',
optimizer=optimizers.Adam(lr=1e-1),
metrics=['accuracy'])
The issue is that your labels are 1 and 2 instead of 0 and 1. Keras will not raise an error when it sees 2, but it is not capable of predicting 2.
Subtract 1 from all your y values. As a side note, it is common in deep learning to use 1 neuron with sigmoid for binary classification (0 or 1) vs 2 classes with softmax. Finally, use binary_crossentropy for the loss for binary classification problems.

Neural Network model with keras python

I have a question about my NN model. I am using keras from python. My training consists of 1000 samples, each with 4320 features. There are 10 categories, and my Y contains numpy arrays of 10 elements with 0 on all the positions except one.
However, my NN doesn't learn from the first epoch and I probably have my model wrong, it's my first attempt of building a NN model and I must have got wrong a couple of things.
Epoch 1/150
1000/1000 [==============================] - 40s 40ms/step - loss: 6.7110 - acc: 0.5796
Epoch 2/150
1000/1000 [==============================] - 39s 39ms/step - loss: 6.7063 - acc: 0.5800
Epoch 3/150
1000/1000 [==============================] - 38s 38ms/step - loss: 6.7063 - acc: 0.5800
Epoch 4/150
1000/1000 [==============================] - 39s 39ms/step - loss: 6.7063 - acc: 0.5800
Epoch 5/150
1000/1000 [==============================] - 38s 38ms/step - loss: 6.7063 - acc: 0.5800
Epoch 6/150
1000/1000 [==============================] - 38s 38ms/step - loss: 6.7063 - acc: 0.5800
Epoch 7/150
1000/1000 [==============================] - 40s 40ms/step - loss: 6.7063 - acc: 0.5800
Epoch 8/150
1000/1000 [==============================] - 39s 39ms/step - loss: 6.7063 - acc: 0.5800
Epoch 9/150
1000/1000 [==============================] - 40s 40ms/step - loss: 6.7063 - acc: 0.5800
And this is part of my NN code:
model = Sequential()
model.add(Dense(4320, input_dim=4320, activation='relu'))
model.add(Dense(50, activation='relu'))
model.add(Dense(10, activation='sigmoid'))
model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])
model.fit(X, Y, epochs=150, batch_size=10)
So, my X is a numpy array of length 1000 that contains other numpy arrays of 4320 elements. My Y is a numpy array of length 1000 that contains other numpy arrays of 10 elements (categories).
Am I doing something wrong or it just can't learn based on this training set? (On 1NN with manhattan distance I'm getting ~80% accuracy on this training set)
L.E.: After I've normalized the data, this is the output of my first 10 epochs:
Epoch 1/150
1000/1000 [==============================] - 41s 41ms/step - loss: 7.9834 - acc: 0.4360
Epoch 2/150
1000/1000 [==============================] - 41s 41ms/step - loss: 7.2943 - acc: 0.5080
Epoch 3/150
1000/1000 [==============================] - 39s 39ms/step - loss: 9.0326 - acc: 0.4070
Epoch 4/150
1000/1000 [==============================] - 39s 39ms/step - loss: 8.7106 - acc: 0.4320
Epoch 5/150
1000/1000 [==============================] - 40s 40ms/step - loss: 7.7547 - acc: 0.4900
Epoch 6/150
1000/1000 [==============================] - 44s 44ms/step - loss: 7.2591 - acc: 0.5270
Epoch 7/150
1000/1000 [==============================] - 42s 42ms/step - loss: 8.5002 - acc: 0.4560
Epoch 8/150
1000/1000 [==============================] - 41s 41ms/step - loss: 9.9525 - acc: 0.3720
Epoch 9/150
1000/1000 [==============================] - 40s 40ms/step - loss: 9.7160 - acc: 0.3920
Epoch 10/150
1000/1000 [==============================] - 39s 39ms/step - loss: 9.3523 - acc: 0.4140
Looks like it starts fluctuating so that seems to be good
It seems like your categories, classes are mutually exclusive since your target arrays are one-hot encoded (ie you never have to predict 2 classes at the same time). In that case, you should use softmax on your last layer to produce a distribution and train using categorical_crossentropy. If fact you can just set your targets as Y = [2,4,0,1] as your category indices and train with sparse_categorical_crossentropy which will save you the time of creating a 2 array of shape (samples, 10).
It seems like you have a lot of features, most likely the performance of your network will depend on how you pre-process your data. For continuous inputs, it's wise to normalise it and for discrete input encode it as one-hot to help the learning.

Categories