I'm currently working on a neural network project, and I need some help understanding the relationships between parameters and the values my neural network is outputing.
My goal is to train a LSTM neural network to detect stress in speech. I'm using a dataset divided into audios of neutral voices and audios of voices under stress. In order to classify which audios contain stress, I'm extracting relevant features from the voices each frame, and then feeding this information into the LSTM neural network.
Since I'm extracting features by frame, the extraction output from audio files with different lenghts also have different lenghts, proportionally to the audio duration. To normalize the neural networks inputs, I'm using a padding technique, which consists in adding zeroes to the end of each extracted features set to meet the biggest set size.
So, for example, if I have 3 audio files, each with these durations: 4, 5, 6 seconds, the extracted features set from the first two audios would be padded with zeroes to meet the third audio extracted set length.
A padded features set looks like this:
[
[9.323346e+00, 9.222625e+00, 8.910659e+00],
[8.751126e+00, 8.432300e+00, 8.046866e+00],
...
[7.439109e+00, 7.380966e+00, 6.092496e+00],
[0, 0, 0],
[0, 0, 0],
[0, 0, 0]
]
The whole dataset dimensions are as follows: (number of audio files) x (number of frames in biggest audio file) x (number of features)
I divided my dataset into a training set, a validation set and a test set. Currently, I have audio files from two public databases, one set with 576 audio files (288 non-stressed, 288 stressed) and other with 240 files (120 non-stressed, 120 stressed).
The following code shows my LSTM implementation using Keras:
N_HIDDEN_CELLS = 100
LEARNING_RATE = 0.00005
BATCH_SIZE = 32
EPOCHS_N = 30
ACTIVATION_FUNCTION = 'softmax'
LOSS_FUNCTION = 'binary_crossentropy'
def create_model(input_shape)
model = keras.Sequential()
model.add(keras.layers.LSTM(N_HIDDEN_CELLS, input_shape=input_shape, return_sequences=True))
model.add(keras.layers.LSTM(N_HIDDEN_CELLS, return_sequences=True))
model.add(keras.layers.LSTM(N_HIDDEN_CELLS, return_sequences=True))
model.add(keras.layers.Dropout(0.3))
model.add(keras.layers.LSTM(2, activation=ACTIVATION_FUNCTION))
return model
def prepare_datasets(data, labels, test_size, validation_size):
X_train, X_test, y_train, y_test = train_test_split(data, labels, test_size=test_size)
X_train, X_validation, y_train, y_validation = train_test_split(X_train, y_train, test_size=validation_size)
return X_train, X_validation, X_test, y_train, y_validation, y_test
X_train, X_validation, X_test, y_train, y_validation, y_test = prepare_datasets(data, labels, 0.25, 0.2)
input_shape = (X_train.shape[1], X_train.shape[2])
model = create_model(input_shape)
optimizer = keras.optimizers.Adam(learning_rate=LEARNING_RATE)
model.compile(optimizer=optimizer, loss=LOSS_FUNCTION, metrics=['accuracy'])
model.summary()
history = model.fit(X_train, y_train, validation_data=(X_validation, y_validation), batch_size=BATCH_SIZE, epochs=EPOCHS_N)
test_loss, test_acc = model.evaluate(X_test, y_test, verbose=2)
After various tests and executions, I'm not so confident about my network performance. At first, the validation loss values were all over the place, varying a lot and not converging at all. With some adjustments to parameters, I ended up with the values in the code above. Still, I'm not that confident, mainly because the validation loss starts to vary after epoch 15 (more or less). In the first epochs, both training and validation losses fall accordingly to expectations, but after some epochs, the training loss keeps falling and the validation loss starts to vary and rise.
Below are two executions of the same network (with the same parameters as the code provided) and same dataset (the one with 576 audio files):
Epoch 1/30
11/11 [==============================] - 5s 194ms/step - loss: 0.8493 - accuracy: 0.4934 - val_loss: 0.8436 - val_accuracy: 0.4943
Epoch 2/30
11/11 [==============================] - 1s 123ms/step - loss: 0.8398 - accuracy: 0.5271 - val_loss: 0.8364 - val_accuracy: 0.4943
Epoch 3/30
11/11 [==============================] - 1s 124ms/step - loss: 0.8291 - accuracy: 0.6015 - val_loss: 0.8277 - val_accuracy: 0.4828
Epoch 4/30
11/11 [==============================] - 1s 128ms/step - loss: 0.8187 - accuracy: 0.6022 - val_loss: 0.8159 - val_accuracy: 0.5402
Epoch 5/30
11/11 [==============================] - 1s 124ms/step - loss: 0.8017 - accuracy: 0.6691 - val_loss: 0.8002 - val_accuracy: 0.5862
Epoch 6/30
11/11 [==============================] - 1s 123ms/step - loss: 0.7754 - accuracy: 0.7081 - val_loss: 0.7750 - val_accuracy: 0.6322
Epoch 7/30
11/11 [==============================] - 1s 124ms/step - loss: 0.7455 - accuracy: 0.7168 - val_loss: 0.7391 - val_accuracy: 0.6092
Epoch 8/30
11/11 [==============================] - 1s 130ms/step - loss: 0.7017 - accuracy: 0.7287 - val_loss: 0.6896 - val_accuracy: 0.6437
Epoch 9/30
11/11 [==============================] - 1s 125ms/step - loss: 0.6519 - accuracy: 0.7210 - val_loss: 0.6311 - val_accuracy: 0.6897
Epoch 10/30
11/11 [==============================] - 1s 129ms/step - loss: 0.5613 - accuracy: 0.7817 - val_loss: 0.5935 - val_accuracy: 0.7356
Epoch 11/30
11/11 [==============================] - 1s 123ms/step - loss: 0.5050 - accuracy: 0.7789 - val_loss: 0.5645 - val_accuracy: 0.7471
Epoch 12/30
11/11 [==============================] - 1s 123ms/step - loss: 0.4612 - accuracy: 0.8098 - val_loss: 0.5127 - val_accuracy: 0.7356
Epoch 13/30
11/11 [==============================] - 1s 127ms/step - loss: 0.4117 - accuracy: 0.8301 - val_loss: 0.4848 - val_accuracy: 0.7931
Epoch 14/30
11/11 [==============================] - 1s 128ms/step - loss: 0.3857 - accuracy: 0.8479 - val_loss: 0.4609 - val_accuracy: 0.7816
Epoch 15/30
11/11 [==============================] - 1s 122ms/step - loss: 0.3392 - accuracy: 0.8724 - val_loss: 0.4467 - val_accuracy: 0.8276
Epoch 16/30
11/11 [==============================] - 1s 118ms/step - loss: 0.3140 - accuracy: 0.8901 - val_loss: 0.4462 - val_accuracy: 0.8161
Epoch 17/30
11/11 [==============================] - 1s 125ms/step - loss: 0.2775 - accuracy: 0.9092 - val_loss: 0.4619 - val_accuracy: 0.8046
Epoch 18/30
11/11 [==============================] - 1s 128ms/step - loss: 0.2963 - accuracy: 0.8873 - val_loss: 0.3995 - val_accuracy: 0.8621
Epoch 19/30
11/11 [==============================] - 1s 122ms/step - loss: 0.2663 - accuracy: 0.9141 - val_loss: 0.4364 - val_accuracy: 0.8276
Epoch 20/30
11/11 [==============================] - 1s 120ms/step - loss: 0.2415 - accuracy: 0.9368 - val_loss: 0.4758 - val_accuracy: 0.8276
Epoch 21/30
11/11 [==============================] - 1s 121ms/step - loss: 0.2209 - accuracy: 0.9297 - val_loss: 0.3855 - val_accuracy: 0.8276
Epoch 22/30
11/11 [==============================] - 1s 121ms/step - loss: 0.1605 - accuracy: 0.9676 - val_loss: 0.3658 - val_accuracy: 0.8621
Epoch 23/30
11/11 [==============================] - 1s 126ms/step - loss: 0.1618 - accuracy: 0.9641 - val_loss: 0.3638 - val_accuracy: 0.8506
Epoch 24/30
11/11 [==============================] - 1s 129ms/step - loss: 0.1309 - accuracy: 0.9728 - val_loss: 0.4450 - val_accuracy: 0.8276
Epoch 25/30
11/11 [==============================] - 1s 125ms/step - loss: 0.2014 - accuracy: 0.9394 - val_loss: 0.3439 - val_accuracy: 0.8621
Epoch 26/30
11/11 [==============================] - 1s 126ms/step - loss: 0.1342 - accuracy: 0.9554 - val_loss: 0.3356 - val_accuracy: 0.8851
Epoch 27/30
11/11 [==============================] - 1s 125ms/step - loss: 0.1555 - accuracy: 0.9618 - val_loss: 0.3486 - val_accuracy: 0.8736
Epoch 28/30
11/11 [==============================] - 1s 124ms/step - loss: 0.1346 - accuracy: 0.9659 - val_loss: 0.3208 - val_accuracy: 0.9080
Epoch 29/30
11/11 [==============================] - 1s 127ms/step - loss: 0.1193 - accuracy: 0.9697 - val_loss: 0.3706 - val_accuracy: 0.8851
Epoch 30/30
11/11 [==============================] - 1s 123ms/step - loss: 0.0836 - accuracy: 0.9777 - val_loss: 0.3623 - val_accuracy: 0.8621
5/5 - 0s - loss: 0.4383 - accuracy: 0.8472
Test accuracy: 0.8472222089767456
Test loss: 0.43826407194137573
1st execution val_loss x train_loss graph
Epoch 1/30
11/11 [==============================] - 5s 190ms/step - loss: 0.8297 - accuracy: 0.5306 - val_loss: 0.8508 - val_accuracy: 0.4138
Epoch 2/30
11/11 [==============================] - 1s 123ms/step - loss: 0.8138 - accuracy: 0.5460 - val_loss: 0.8355 - val_accuracy: 0.4713
Epoch 3/30
11/11 [==============================] - 1s 120ms/step - loss: 0.8082 - accuracy: 0.5384 - val_loss: 0.8145 - val_accuracy: 0.5402
Epoch 4/30
11/11 [==============================] - 1s 118ms/step - loss: 0.7997 - accuracy: 0.5799 - val_loss: 0.7911 - val_accuracy: 0.5517
Epoch 5/30
11/11 [==============================] - 1s 117ms/step - loss: 0.7752 - accuracy: 0.6585 - val_loss: 0.7654 - val_accuracy: 0.5862
Epoch 6/30
11/11 [==============================] - 1s 125ms/step - loss: 0.7527 - accuracy: 0.6609 - val_loss: 0.7289 - val_accuracy: 0.6437
Epoch 7/30
11/11 [==============================] - 1s 121ms/step - loss: 0.7129 - accuracy: 0.7432 - val_loss: 0.6790 - val_accuracy: 0.6782
Epoch 8/30
11/11 [==============================] - 1s 125ms/step - loss: 0.6570 - accuracy: 0.7707 - val_loss: 0.6107 - val_accuracy: 0.7356
Epoch 9/30
11/11 [==============================] - 1s 125ms/step - loss: 0.6112 - accuracy: 0.7513 - val_loss: 0.5529 - val_accuracy: 0.7586
Epoch 10/30
11/11 [==============================] - 1s 129ms/step - loss: 0.5339 - accuracy: 0.8026 - val_loss: 0.4895 - val_accuracy: 0.7816
Epoch 11/30
11/11 [==============================] - 1s 120ms/step - loss: 0.4720 - accuracy: 0.8189 - val_loss: 0.4579 - val_accuracy: 0.8046
Epoch 12/30
11/11 [==============================] - 1s 121ms/step - loss: 0.4332 - accuracy: 0.8527 - val_loss: 0.4169 - val_accuracy: 0.8046
Epoch 13/30
11/11 [==============================] - 1s 122ms/step - loss: 0.3976 - accuracy: 0.8568 - val_loss: 0.3850 - val_accuracy: 0.7931
Epoch 14/30
11/11 [==============================] - 1s 124ms/step - loss: 0.3489 - accuracy: 0.8726 - val_loss: 0.3753 - val_accuracy: 0.8046
Epoch 15/30
11/11 [==============================] - 1s 124ms/step - loss: 0.3088 - accuracy: 0.9020 - val_loss: 0.3562 - val_accuracy: 0.8161
Epoch 16/30
11/11 [==============================] - 1s 124ms/step - loss: 0.3489 - accuracy: 0.8745 - val_loss: 0.3501 - val_accuracy: 0.8391
Epoch 17/30
11/11 [==============================] - 1s 130ms/step - loss: 0.2725 - accuracy: 0.9240 - val_loss: 0.3436 - val_accuracy: 0.8506
Epoch 18/30
11/11 [==============================] - 1s 121ms/step - loss: 0.3494 - accuracy: 0.8764 - val_loss: 0.3516 - val_accuracy: 0.8506
Epoch 19/30
11/11 [==============================] - 1s 119ms/step - loss: 0.2553 - accuracy: 0.9243 - val_loss: 0.3413 - val_accuracy: 0.8391
Epoch 20/30
11/11 [==============================] - 1s 122ms/step - loss: 0.2723 - accuracy: 0.9092 - val_loss: 0.3258 - val_accuracy: 0.8621
Epoch 21/30
11/11 [==============================] - 1s 121ms/step - loss: 0.2600 - accuracy: 0.9306 - val_loss: 0.3257 - val_accuracy: 0.8506
Epoch 22/30
11/11 [==============================] - 1s 126ms/step - loss: 0.2406 - accuracy: 0.9411 - val_loss: 0.3203 - val_accuracy: 0.8966
Epoch 23/30
11/11 [==============================] - 1s 127ms/step - loss: 0.1892 - accuracy: 0.9577 - val_loss: 0.3191 - val_accuracy: 0.8851
Epoch 24/30
11/11 [==============================] - 1s 127ms/step - loss: 0.1869 - accuracy: 0.9594 - val_loss: 0.3246 - val_accuracy: 0.8621
Epoch 25/30
11/11 [==============================] - 1s 122ms/step - loss: 0.1898 - accuracy: 0.9487 - val_loss: 0.3217 - val_accuracy: 0.8851
Epoch 26/30
11/11 [==============================] - 1s 125ms/step - loss: 0.1731 - accuracy: 0.9523 - val_loss: 0.3280 - val_accuracy: 0.8506
Epoch 27/30
11/11 [==============================] - 1s 128ms/step - loss: 0.1445 - accuracy: 0.9687 - val_loss: 0.3213 - val_accuracy: 0.8851
Epoch 28/30
11/11 [==============================] - 1s 117ms/step - loss: 0.1441 - accuracy: 0.9718 - val_loss: 0.3212 - val_accuracy: 0.8621
Epoch 29/30
11/11 [==============================] - 1s 124ms/step - loss: 0.1250 - accuracy: 0.9762 - val_loss: 0.3232 - val_accuracy: 0.8851
Epoch 30/30
11/11 [==============================] - 1s 123ms/step - loss: 0.1460 - accuracy: 0.9687 - val_loss: 0.3218 - val_accuracy: 0.8736
5/5 - 0s - loss: 0.3297 - accuracy: 0.8889
Test accuracy: 0.8888888955116272
Test loss: 0.32971107959747314
2nd execution val_loss x train_loss graph
Some additional information:
My labels are hot encoded.
Frame step is 0.05s.
Frame size is 0.125s.
When running this configuration with the smaller dataset, I get a slightly different behaviour. The loss value falls more evenly, but kind of slowly. I tried increasing the epochs number, but after the 30th epoch the validation loss started to vary and rise as well.
My questions are:
What could be causing this validation loss problem?
What does it mean when a model has a high loss rate, but its accuracy remains ok?
I read about binary cross entropy but I don't know if I understand what the loss value means in my tests, could someone help me understanding these values?
Could this padding strategy be affecting the network performance?
Is my input data and its dimensions coherent considering LSTM definitions?
Could this be related to my dataset size?
What would be an acceptable validation loss rate?
Your validation loss being much higher than your training loss usually implies overfitting. Note that your val_loss isn't really high, it's just higher than the training loss. The validation accuracy isn't bad either, just much lower than on the training data which has effectively been memorized by your network.
Basically you need to reduce the strength of the model so it can generalize to the complexity of the problem at hand. Use more dropout and fewer parameters/layers.
Related
I am trying to train model in three machines using tensorflow MultiWorkerMirroredStrategy. The script is based on the tensorflow tutorial:Multi-worker training with Keras(https://www.tensorflow.org/tutorials/distribute/multi_worker_with_keras#dataset_sharding_and_batch_size):
import tensorflow_datasets as tfds
import tensorflow as tf
tfds.disable_progress_bar()
import os
import json
strategy = tf.distribute.MultiWorkerMirroredStrategy()
BUFFER_SIZE = 10000
BATCH_SIZE = 64
def make_datasets_unbatched():
# scale MNIST data from (0, 255] to (0., 1.]
def scale(image, label):
image = tf.cast(image, tf.float32)
image /= 255
return image, label
# data download to /home/pzs/tensorflow_datasets/mnist/
datasets, info = tfds.load(name='mnist',
with_info=True,
as_supervised=True)
return datasets['train'].map(scale).cache().shuffle(BUFFER_SIZE)
def build_and_compile_cnn_model():
model = tf.keras.Sequential([
tf.keras.layers.Conv2D(32, 3, activation='relu', input_shape=(28, 28, 1)),
tf.keras.layers.MaxPooling2D(),
tf.keras.layers.Flatten(),
tf.keras.layers.Dense(64, activation='relu'),
tf.keras.layers.Dense(10)
])
model.compile(
loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),
optimizer=tf.keras.optimizers.SGD(learning_rate=0.001),
metrics=['accuracy'])
return model
NUM_WORKERS = 3
GLOBAL_BATCH_SIZE = 64 * NUM_WORKERS
train_datasets = make_datasets_unbatched().batch(GLOBAL_BATCH_SIZE)
options = tf.data.Options()
options.experimental_distribute.auto_shard_policy = tf.data.experimental.AutoShardPolicy.OFF
train_datasets = make_datasets_unbatched().batch(BATCH_SIZE)
train_datasets = train_datasets.with_options(options)
with strategy.scope():
multi_worker_model = build_and_compile_cnn_model()
multi_worker_model.fit(x=train_datasets, epochs=30, steps_per_epoch=5)
I run this script separately on tree node3:
on node 1:
TF_CONFIG='{"cluster": {"worker": ["192.168.4.36:12346", "192.168.4.83:12346", "192.168.4.83:12346"]}, "task": {"index": 0, "type": "worker"}}' python3 multi_worker_with_keras.py
on node 2:
TF_CONFIG='{"cluster": {"worker": ["192.168.4.36:12346", "192.168.4.83:12346", "192.168.4.83:12346"]}, "task": {"index": 1, "type": "worker"}}' python3 multi_worker_with_keras.py
on node 3:
TF_CONFIG='{"cluster": {"worker": ["192.168.4.36:12346", "192.168.4.83:12346", "192.168.4.83:12346"]}, "task": {"index": 2, "type": "worker"}}' python3 multi_worker_with_keras.py
and the results of training error and accuracy are:
Epoch 1/30
2022-02-16 11:52:25.060362: I tensorflow/stream_executor/cuda/cuda_dnn.cc:366] Loaded cuDNN version 8201
5/5 [==============================] - 7s 195ms/step - loss: 2.3010 - accuracy: 0.0719
Epoch 2/30
5/5 [==============================] - 1s 181ms/step - loss: 2.2984 - accuracy: 0.0688
Epoch 3/30
5/5 [==============================] - 1s 182ms/step - loss: 2.2993 - accuracy: 0.0781
Epoch 4/30
5/5 [==============================] - 1s 182ms/step - loss: 2.2917 - accuracy: 0.0594
Epoch 5/30
5/5 [==============================] - 1s 182ms/step - loss: 2.2987 - accuracy: 0.0969
Epoch 6/30
5/5 [==============================] - 1s 183ms/step - loss: 2.2992 - accuracy: 0.0906
Epoch 7/30
5/5 [==============================] - 1s 181ms/step - loss: 2.2978 - accuracy: 0.1000
Epoch 8/30
5/5 [==============================] - 1s 183ms/step - loss: 2.2887 - accuracy: 0.0969
Epoch 9/30
5/5 [==============================] - 1s 182ms/step - loss: 2.2887 - accuracy: 0.0969
Epoch 10/30
5/5 [==============================] - 1s 183ms/step - loss: 2.2930 - accuracy: 0.0844
Epoch 11/30
5/5 [==============================] - 1s 184ms/step - loss: 2.2905 - accuracy: 0.1000
Epoch 12/30
5/5 [==============================] - 1s 184ms/step - loss: 2.2884 - accuracy: 0.0812
Epoch 13/30
5/5 [==============================] - 1s 186ms/step - loss: 2.2837 - accuracy: 0.1250
Epoch 14/30
5/5 [==============================] - 1s 189ms/step - loss: 2.2842 - accuracy: 0.1094
Epoch 15/30
5/5 [==============================] - 1s 190ms/step - loss: 2.2856 - accuracy: 0.0750
Epoch 16/30
5/5 [==============================] - 1s 192ms/step - loss: 2.2911 - accuracy: 0.0719
Epoch 17/30
5/5 [==============================] - 1s 188ms/step - loss: 2.2805 - accuracy: 0.1031
Epoch 18/30
5/5 [==============================] - 1s 187ms/step - loss: 2.2800 - accuracy: 0.1219
Epoch 19/30
5/5 [==============================] - 1s 190ms/step - loss: 2.2799 - accuracy: 0.1063
Epoch 20/30
5/5 [==============================] - 1s 192ms/step - loss: 2.2769 - accuracy: 0.1187
Epoch 21/30
5/5 [==============================] - 1s 193ms/step - loss: 2.2768 - accuracy: 0.1344
Epoch 22/30
5/5 [==============================] - 1s 190ms/step - loss: 2.2754 - accuracy: 0.1187
Epoch 23/30
5/5 [==============================] - 1s 190ms/step - loss: 2.2821 - accuracy: 0.1187
Epoch 24/30
5/5 [==============================] - 1s 188ms/step - loss: 2.2832 - accuracy: 0.0844
Epoch 25/30
5/5 [==============================] - 1s 190ms/step - loss: 2.2793 - accuracy: 0.1125
Epoch 26/30
5/5 [==============================] - 1s 191ms/step - loss: 2.2762 - accuracy: 0.1406
Epoch 27/30
5/5 [==============================] - 1s 194ms/step - loss: 2.2696 - accuracy: 0.1344
Epoch 28/30
5/5 [==============================] - 1s 192ms/step - loss: 2.2717 - accuracy: 0.1406
Epoch 29/30
5/5 [==============================] - 1s 191ms/step - loss: 2.2680 - accuracy: 0.1500
Epoch 30/30
5/5 [==============================] - 1s 193ms/step - loss: 2.2696 - accuracy: 0.1500
all results are exactly the same for 3 nodes.
my question is:
When using tf.distribute.MultiWorkerMirroredStrategy to train model among multiple machines, each process does forward and backward propagation independently using different slice of a batch training data, why the training errors are all the same for the corresponding epoch in 3 nodes? I try to run a different script and found the same case.
This is expected. The metric values would be allreduced in fit method.
https://github.com/tensorflow/tensorflow/issues/39343#issuecomment-627008557
CNN is:
input: 2 image 128x128: input1 and input2
output: CNN return a Trasformation Matrix:
outLayer = Layers.Dense(3 * 2, activation=activations.linear,kernel_initializer="zeos",bias_initializer=output_bias)(d2)
After I:
append STN (https://github.com/kevinzakka/spatial-transformer-network) for translate and rotate input1.
The complete network is CNN with STN. Return an image (128x128) with 32 batch
My loss is MSE between input2 and output (CNN): target error registration. MSE between 2 image (input2 and output)
The network must find the transformation matrix because input2 = output
Can I use STN only for translate and rotate input1? Because I have problem with val_loss (use tensorflow):
Epoch 1/90
30/30 [==============================] - 81s 3s/step - loss: 16.0190 - val_loss: 24.3248
Epoch 2/90
30/30 [==============================] - 79s 3s/step - loss: 13.9868 - val_loss: 21.4465
Epoch 3/90
30/30 [==============================] - 73s 2s/step - loss: 13.2970 - val_loss: 21.3151
Epoch 4/90
30/30 [==============================] - 69s 2s/step - loss: 12.9244 - val_loss: 21.6154
Epoch 5/90
30/30 [==============================] - 67s 2s/step - loss: 12.6868 - val_loss: 20.0113
Epoch 6/90
30/30 [==============================] - 66s 2s/step - loss: 12.4998 - val_loss: 20.8911
Epoch 7/90
30/30 [==============================] - 69s 2s/step - loss: 12.3066 - val_loss: 21.4276
Epoch 8/90
30/30 [==============================] - 67s 2s/step - loss: 12.1034 - val_loss: 21.3593
Epoch 9/90
30/30 [==============================] - 69s 2s/step - loss: 11.7645 - val_loss: 20.8941
Epoch 10/90
30/30 [==============================] - 67s 2s/step - loss: 11.6544 - val_loss: 20.4768
Epoch 11/90
30/30 [==============================] - 70s 2s/step - loss: 11.5483 - val_loss: 21.7420
Epoch 12/90
30/30 [==============================] - 68s 2s/step - loss: 11.5680 - val_loss: 19.4531
Here is my code and result of the training.
batch_size = 100
epochs = 50
yale_history = yale_classifier.fit(x_train, y_train_oh,batch_size=batch_size,epochs=epochs,validation_data=(x_train,y_train_oh))
Epoch 1/50
20/20 [==============================] - 32s 2s/step - loss: 3.9801 - accuracy: 0.2071 - val_loss: 3.6919 - val_accuracy: 0.0245
Epoch 2/50
20/20 [==============================] - 30s 2s/step - loss: 1.2557 - accuracy: 0.6847 - val_loss: 4.1914 - val_accuracy: 0.0245
Epoch 3/50
20/20 [==============================] - 30s 2s/step - loss: 0.4408 - accuracy: 0.8954 - val_loss: 4.6284 - val_accuracy: 0.0245
Epoch 4/50
20/20 [==============================] - 30s 2s/step - loss: 0.1822 - accuracy: 0.9592 - val_loss: 4.9481 - val_accuracy: 0.0398
Epoch 5/50
20/20 [==============================] - 30s 2s/step - loss: 0.1252 - accuracy: 0.9760 - val_loss: 5.3728 - val_accuracy: 0.0276
Epoch 6/50
20/20 [==============================] - 30s 2s/step - loss: 0.0927 - accuracy: 0.9816 - val_loss: 5.7009 - val_accuracy: 0.0260
Epoch 7/50
20/20 [==============================] - 30s 2s/step - loss: 0.0858 - accuracy: 0.9837 - val_loss: 6.0049 - val_accuracy: 0.0260
Epoch 8/50
20/20 [==============================] - 30s 2s/step - loss: 0.0646 - accuracy: 0.9867 - val_loss: 6.3786 - val_accuracy: 0.0260
Epoch 9/50
20/20 [==============================] - 30s 2s/step - loss: 0.0489 - accuracy: 0.9898 - val_loss: 6.5156 - val_accuracy: 0.0260
You can see that I also used the training data as the validation data. This is weird that the training loss is not the same as the validation loss. Further, when I evaluated it, seem like my model was not trained at all as follow.
yale_classifier.evaluate(x_train, y_train_oh)
62/62 [==============================] - 6s 96ms/step - loss: 7.1123 - accuracy: 0.0260
[7.112329483032227, 0.026020407676696777]
Do you have any recommened to solve this problem ?
I am fairly new to deep learning and right now am trying to predict consumer choices based on EEG data. The total dataset consists of 1045 EEG recordings each with a corresponding label, indicating Like or Dislike for a product. Classes are distributed as follows (44% Likes and 56% Dislikes). I read that Convolutional Neural Networks are suitable to work with raw EEG data so I tried to implement a network based on keras with the following structure:
from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(full_data, target, test_size=0.20, random_state=42)
y_train = np.asarray(y_train).astype('float32').reshape((-1,1))
y_test = np.asarray(y_test).astype('float32').reshape((-1,1))
# X_train.shape = ((836, 512, 14))
# y_train.shape = ((836, 1))
from keras.optimizers import Adam
from keras.optimizers import SGD
from keras.layers import MaxPooling1D
model = Sequential()
model.add(Conv1D(16, kernel_size=3, activation="relu", input_shape=(512,14)))
model.add(MaxPooling1D())
model.add(Conv1D(8, kernel_size=3, activation="relu"))
model.add(MaxPooling1D())
model.add(Flatten())
model.add(Dense(1, activation="sigmoid"))
model.compile(optimizer=Adam(lr = 0.001), loss='binary_crossentropy', metrics=['accuracy'])
model.fit(X_train, y_train, validation_data=(X_test, y_test), epochs=20, batch_size = 64)
When I fit the model however the validation accuracy does not change at all with the following output:
Epoch 1/20
14/14 [==============================] - 0s 32ms/step - loss: 292.6353 - accuracy: 0.5383 - val_loss: 0.7884 - val_accuracy: 0.5407
Epoch 2/20
14/14 [==============================] - 0s 7ms/step - loss: 1.3748 - accuracy: 0.5598 - val_loss: 0.8860 - val_accuracy: 0.5502
Epoch 3/20
14/14 [==============================] - 0s 6ms/step - loss: 1.0537 - accuracy: 0.5598 - val_loss: 0.7629 - val_accuracy: 0.5455
Epoch 4/20
14/14 [==============================] - 0s 6ms/step - loss: 0.8827 - accuracy: 0.5598 - val_loss: 0.7010 - val_accuracy: 0.5455
Epoch 5/20
14/14 [==============================] - 0s 6ms/step - loss: 0.7988 - accuracy: 0.5598 - val_loss: 0.8689 - val_accuracy: 0.5407
Epoch 6/20
14/14 [==============================] - 0s 6ms/step - loss: 1.0221 - accuracy: 0.5610 - val_loss: 0.6961 - val_accuracy: 0.5455
Epoch 7/20
14/14 [==============================] - 0s 6ms/step - loss: 0.7415 - accuracy: 0.5598 - val_loss: 0.6945 - val_accuracy: 0.5455
Epoch 8/20
14/14 [==============================] - 0s 6ms/step - loss: 0.7381 - accuracy: 0.5574 - val_loss: 0.7761 - val_accuracy: 0.5455
Epoch 9/20
14/14 [==============================] - 0s 6ms/step - loss: 0.7326 - accuracy: 0.5598 - val_loss: 0.6926 - val_accuracy: 0.5455
Epoch 10/20
14/14 [==============================] - 0s 6ms/step - loss: 0.7338 - accuracy: 0.5598 - val_loss: 0.6917 - val_accuracy: 0.5455
Epoch 11/20
14/14 [==============================] - 0s 7ms/step - loss: 0.7203 - accuracy: 0.5610 - val_loss: 0.6916 - val_accuracy: 0.5455
Epoch 12/20
14/14 [==============================] - 0s 6ms/step - loss: 0.7192 - accuracy: 0.5610 - val_loss: 0.6914 - val_accuracy: 0.5455
Epoch 13/20
14/14 [==============================] - 0s 6ms/step - loss: 0.7174 - accuracy: 0.5610 - val_loss: 0.6912 - val_accuracy: 0.5455
Epoch 14/20
14/14 [==============================] - 0s 6ms/step - loss: 0.7155 - accuracy: 0.5610 - val_loss: 0.6911 - val_accuracy: 0.5455
Epoch 15/20
14/14 [==============================] - 0s 6ms/step - loss: 0.7143 - accuracy: 0.5610 - val_loss: 0.6910 - val_accuracy: 0.5455
Epoch 16/20
14/14 [==============================] - 0s 6ms/step - loss: 0.7129 - accuracy: 0.5610 - val_loss: 0.6909 - val_accuracy: 0.5455
Epoch 17/20
14/14 [==============================] - 0s 6ms/step - loss: 0.7114 - accuracy: 0.5610 - val_loss: 0.6907 - val_accuracy: 0.5455
Epoch 18/20
14/14 [==============================] - 0s 6ms/step - loss: 0.7103 - accuracy: 0.5610 - val_loss: 0.6906 - val_accuracy: 0.5455
Epoch 19/20
14/14 [==============================] - 0s 6ms/step - loss: 0.7088 - accuracy: 0.5610 - val_loss: 0.6906 - val_accuracy: 0.5455
Epoch 20/20
14/14 [==============================] - 0s 6ms/step - loss: 0.7075 - accuracy: 0.5610 - val_loss: 0.6905 - val_accuracy: 0.5455
Thanks in advance for any insights!
The phenomenon you run into is called underfitting. This happens when the amount our quality of your training data is insufficient, or your network architecture is too small and not capable to learn the problem.
Try normalizing your input data and experiment with different network architectures, learning rates and activation functions.
As #Muhammad Shahzad stated in his comment, adding some Dense Layers after flatting would be a concrete architecture adaption you should try.
You can also increase the epoch and must increase the data set. And you also can use-
train_datagen= ImageDataGenerator(
rescale=1./255,
shear_range=0.2,
zoom_range=0.2,
horizontal_flip=True,
vertical_flip = True,
channel_shift_range=0.2,
fill_mode='nearest'
)
for feeding the model more data and I hope you can increase the validation_accuracy.
I am doing a Binary classification of IMDB movie review data into Positive or Negative Sentiment.
I have 25K movie reviews and corresponding label.
Preprocessing:
Removed the stop words and split the data into 70:30 training and test. So 17.5K training and 7k test. 17.5k training has been further divided into 14K train and 3.5 k validation dataset as used in keras.model.fit method
Each processed movie review has been converted to TF-IDF vector using Keras text processing module.
Here is my Fully Connected Architecture I used in Keras Dense class
def model_param(self):
""" Method to do deep learning
"""
from keras.models import Sequential
from keras.layers import Dense, Dropout, Activation
from keras.optimizers import SGD
from keras import regularizers
self.model = Sequential()
# Dense(64) is a fully-connected layer with 64 hidden units.
# in the first layer, you must specify the expected input data shape:
# here, 20-dimensional vectors.
self.model.add(Dense(32, activation='relu', input_dim=self.x_train_std.shape[1]))
self.model.add(Dropout(0.5))
#self.model.add(Dense(60, activation='relu'))
#self.model.add(Dropout(0.5))
self.model.add(Dense(1, activation='sigmoid'))
sgd = SGD(lr=0.01, decay=1e-6, momentum=0.9, nesterov=True)
self.model.compile(loss='binary_crossentropy',
optimizer=sgd,
metrics=['accuracy'])
def fit(self):
""" Training the deep learning network on the training data
"""
self.model.fit(self.x_train_std, self.y_train,validation_split=0.20,
epochs=50,
batch_size=128)
As you see, I tried first without Dropout and as usual I got training accuracy as 1.0 but validation was poor as overfitting was happening. So I added Dropout to prevent overfitting
However inspite of trying multiple dropout ratio, adding another layer with different no. of units in it, changing learning rate I am still getting overfitting on validation dataset. Gets stuck at 85% while training keeps increasing to 99% and so on. Even changed the Epochs from 10 to 50
What could be going wrong here
Train on 14000 samples, validate on 3500 samples
Epoch 1/50
14000/14000 [==============================] - 0s - loss: 0.5684 - acc: 0.7034 - val_loss: 0.3794 - val_acc: 0.8431
Epoch 2/50
14000/14000 [==============================] - 0s - loss: 0.3630 - acc: 0.8388 - val_loss: 0.3304 - val_acc: 0.8549
Epoch 3/50
14000/14000 [==============================] - 0s - loss: 0.2977 - acc: 0.8749 - val_loss: 0.3271 - val_acc: 0.8591
Epoch 4/50
14000/14000 [==============================] - 0s - loss: 0.2490 - acc: 0.8991 - val_loss: 0.3302 - val_acc: 0.8580
Epoch 5/50
14000/14000 [==============================] - 0s - loss: 0.2251 - acc: 0.9086 - val_loss: 0.3388 - val_acc: 0.8546
Epoch 6/50
14000/14000 [==============================] - 0s - loss: 0.2021 - acc: 0.9189 - val_loss: 0.3532 - val_acc: 0.8523
Epoch 7/50
14000/14000 [==============================] - 0s - loss: 0.1797 - acc: 0.9286 - val_loss: 0.3670 - val_acc: 0.8529
Epoch 8/50
14000/14000 [==============================] - 0s - loss: 0.1611 - acc: 0.9350 - val_loss: 0.3860 - val_acc: 0.8543
Epoch 9/50
14000/14000 [==============================] - 0s - loss: 0.1427 - acc: 0.9437 - val_loss: 0.4077 - val_acc: 0.8529
Epoch 10/50
14000/14000 [==============================] - 0s - loss: 0.1344 - acc: 0.9476 - val_loss: 0.4234 - val_acc: 0.8526
Epoch 11/50
14000/14000 [==============================] - 0s - loss: 0.1222 - acc: 0.9534 - val_loss: 0.4473 - val_acc: 0.8506
Epoch 12/50
14000/14000 [==============================] - 0s - loss: 0.1131 - acc: 0.9546 - val_loss: 0.4718 - val_acc: 0.8497
Epoch 13/50
14000/14000 [==============================] - 0s - loss: 0.1079 - acc: 0.9559 - val_loss: 0.4818 - val_acc: 0.8526
Epoch 14/50
14000/14000 [==============================] - 0s - loss: 0.0954 - acc: 0.9630 - val_loss: 0.5057 - val_acc: 0.8494
Epoch 15/50
14000/14000 [==============================] - 0s - loss: 0.0906 - acc: 0.9636 - val_loss: 0.5229 - val_acc: 0.8557
Epoch 16/50
14000/14000 [==============================] - 0s - loss: 0.0896 - acc: 0.9657 - val_loss: 0.5387 - val_acc: 0.8497
Epoch 17/50
14000/14000 [==============================] - 0s - loss: 0.0816 - acc: 0.9666 - val_loss: 0.5579 - val_acc: 0.8463
Epoch 18/50
14000/14000 [==============================] - 0s - loss: 0.0762 - acc: 0.9709 - val_loss: 0.5704 - val_acc: 0.8491
Epoch 19/50
14000/14000 [==============================] - 0s - loss: 0.0718 - acc: 0.9723 - val_loss: 0.5834 - val_acc: 0.8454
Epoch 20/50
14000/14000 [==============================] - 0s - loss: 0.0633 - acc: 0.9752 - val_loss: 0.6032 - val_acc: 0.8494
Epoch 21/50
14000/14000 [==============================] - 0s - loss: 0.0687 - acc: 0.9724 - val_loss: 0.6181 - val_acc: 0.8480
Epoch 22/50
14000/14000 [==============================] - 0s - loss: 0.0614 - acc: 0.9762 - val_loss: 0.6280 - val_acc: 0.8503
Epoch 23/50
14000/14000 [==============================] - 0s - loss: 0.0620 - acc: 0.9756 - val_loss: 0.6407 - val_acc: 0.8500
Epoch 24/50
14000/14000 [==============================] - 0s - loss: 0.0536 - acc: 0.9794 - val_loss: 0.6563 - val_acc: 0.8511
Epoch 25/50
14000/14000 [==============================] - 0s - loss: 0.0538 - acc: 0.9791 - val_loss: 0.6709 - val_acc: 0.8500
Epoch 26/50
14000/14000 [==============================] - 0s - loss: 0.0507 - acc: 0.9807 - val_loss: 0.6869 - val_acc: 0.8491
Epoch 27/50
14000/14000 [==============================] - 0s - loss: 0.0528 - acc: 0.9794 - val_loss: 0.7002 - val_acc: 0.8483
Epoch 28/50
14000/14000 [==============================] - 0s - loss: 0.0465 - acc: 0.9810 - val_loss: 0.7083 - val_acc: 0.8469
Epoch 29/50
14000/14000 [==============================] - 0s - loss: 0.0504 - acc: 0.9796 - val_loss: 0.7153 - val_acc: 0.8497
Epoch 30/50
14000/14000 [==============================] - 0s - loss: 0.0477 - acc: 0.9819 - val_loss: 0.7232 - val_acc: 0.8480
Epoch 31/50
14000/14000 [==============================] - 0s - loss: 0.0475 - acc: 0.9819 - val_loss: 0.7343 - val_acc: 0.8469
Epoch 32/50
14000/14000 [==============================] - 0s - loss: 0.0459 - acc: 0.9819 - val_loss: 0.7352 - val_acc: 0.8500
Epoch 33/50
14000/14000 [==============================] - 0s - loss: 0.0426 - acc: 0.9807 - val_loss: 0.7429 - val_acc: 0.8511
Epoch 34/50
14000/14000 [==============================] - 0s - loss: 0.0396 - acc: 0.9846 - val_loss: 0.7576 - val_acc: 0.8477
Epoch 35/50
14000/14000 [==============================] - 0s - loss: 0.0420 - acc: 0.9836 - val_loss: 0.7603 - val_acc: 0.8506
Epoch 36/50
14000/14000 [==============================] - 0s - loss: 0.0359 - acc: 0.9856 - val_loss: 0.7683 - val_acc: 0.8497
Epoch 37/50
14000/14000 [==============================] - 0s - loss: 0.0377 - acc: 0.9849 - val_loss: 0.7823 - val_acc: 0.8520
Epoch 38/50
14000/14000 [==============================] - 0s - loss: 0.0352 - acc: 0.9861 - val_loss: 0.7912 - val_acc: 0.8500
Epoch 39/50
14000/14000 [==============================] - 0s - loss: 0.0390 - acc: 0.9845 - val_loss: 0.8025 - val_acc: 0.8489
Epoch 40/50
14000/14000 [==============================] - 0s - loss: 0.0371 - acc: 0.9853 - val_loss: 0.8128 - val_acc: 0.8494
Epoch 41/50
14000/14000 [==============================] - 0s - loss: 0.0367 - acc: 0.9848 - val_loss: 0.8184 - val_acc: 0.8503
Epoch 42/50
14000/14000 [==============================] - 0s - loss: 0.0331 - acc: 0.9871 - val_loss: 0.8264 - val_acc: 0.8500
Epoch 43/50
14000/14000 [==============================] - 0s - loss: 0.0338 - acc: 0.9871 - val_loss: 0.8332 - val_acc: 0.8483
Epoch 44/50