Menu
Problem Statement
I’m training a many-to-many LSTM in keras with tensorflow backend (tf version 1.13.1) on tagged text sequences to predict the tag of each element in the sequence using pretrained GloVe embeddings. My training regime involves mini batch stochastic gradient descent, with each mini batch matrix zero-padded column-wise to ensure equal length input to the network.
Crucially, because of custom constrains on my mini batches due to the nature of the task and the data, I am not using the keras embedding layer. My goal is to implement a masking mechanism for my zero-padded cells to ensure the loss computation does not spuriously treat these cells as genuine data points.
Approach
As explained in the keras documentation, keras has three ways in which a masking layer can be set up:
keras.layers.Embedding layer with mask_zero
set to True.keras.layers.Masking layer;Because I am not using an embedding layer to encode my data for training, option (1) with a masked embedding layer is not available to me. So instead, I chose (2) and added a masking layer right after initializing my model. This change, however, does not seem to have had an effect. In fact, not only has the accuracy of my model not improved, at the prediction stage the model still generates zero predictions. Why does my masking layer not mask zero-padded cells? Could it have to do with the fact that in my dense layer I'm specifying 3 classes rather than 2 (thus including 0 as a separate class)?
Limitations of Existing Resources
Similar questions have been asked and answered, but I wasn't able to use them to resolve my issue. While this post received no direct response, a linked post mentioned in a comment focuses on how to preprocess data to assign mask value, which is uncontroversial here. The masking layer initializtion, however, is identical to the one used here. This post mentions the same problem - a masking layer has no effect on performance - and the answer defines the masking layer in the same way as I do, but again focuses on converting specific values to mask values. Finally, the answer in this post provides the same layer initialization without elaborating further.
Toy Data Generation
To reproduce my problem, I have generated a toy 10-batch dataset with two classes (1,2). A batch is a variable-length sequence post-padded with zeros to a maximum length of 20 embeddings, with each embedding vector consisting of 5 cells, so input_shape=(20,5). Embedding values for the two classes were generated from different but partially-overlapping truncated normal distributions to create a learnable but not trivial problem for the network. I've included the toy data below so you can reproduce the problem.
import pandas as pd
from keras.models import Sequential
from keras.layers import LSTM, Dense, TimeDistributed, Bidirectional, Dropout, Masking
from keras import optimizers
# *** model initialization ***
model = Sequential()
model.add(Masking(mask_value=0., input_shape=(20, 5))) # <- masking layer here
model.add(Bidirectional(LSTM(20, return_sequences=True), input_shape=(20, 5)))
model.add(Dropout(0.2))
model.add(TimeDistributed(Dense(3, activation='sigmoid')))
sgd = optimizers.SGD(lr=0.01, decay=1e-6, momentum=0.9, nesterov=True)
model.compile(loss='sparse_categorical_crossentropy', optimizer=sgd, metrics=['mse'])
# *** model training ***
for epoch in range(10):
for X,y in data_train:
X = X.reshape(1, 20, 5)
y = y.reshape(1, 20, 1)
history = model.fit(X, y, epochs=1, batch_size=20, verbose=0)
# *** model prediction ***
preds = pd.DataFrame(columns=['true', 'pred'])
for index, (X,y) in enumerate(data_test):
X = X.reshape(1, 20, 5)
y = y.reshape(1, 20, 1)
y_pred = model.predict_classes(X, verbose=0)
df = pd.DataFrame(columns=['true', 'pred'])
df['true'] = [y[0, i][0] for i in range(20)]
df['pred'] = [y_pred[0, i] for i in range(20)]
preds = preds.append(df, ignore_index=True)
# convert true labels to int & drop padded rows (where y_true=0)
preds['true'] = [int(label) for label in preds['true']]
preds = preds[preds['true']!=0]
This is the summary of the model with masking:
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
masking_2 (Masking) (None, 20, 5) 0
_________________________________________________________________
bidirectional_4 (Bidirection (None, 20, 40) 4160
_________________________________________________________________
dropout_4 (Dropout) (None, 20, 40) 0
_________________________________________________________________
time_distributed_4 (TimeDist (None, 20, 3) 123
=================================================================
Total params: 4,283
Trainable params: 4,283
Non-trainable params: 0
I trained one model with and one without the masking layer and calculated accuracy using:
np.round(sum(preds['true']==preds['pred'])/len(preds)*100,1)
I got 53.3% accuracy for the model without masking and 33.3% for the model with masking. More surprisingly, I kept on getting zero as a predicted label in both models. Why does the masking layer fail to ignore zero-padded cells?
Data for reproducing issue:
data_train = list(zip(X_batches_train, y_batches_train))
data_test = list(zip(X_batches_test, y_batches_test))
X_batches_train
[array([[-1.00612917, 1.47313952, 2.68021318, 1.54875809, 0.98385996,
1.49465265, 0.60429106, 1.12396908, -0.24041602, 1.77266187,
0.1961381 , 1.28019637, 1.78803092, 2.05151245, 0.93606708,
0.51554755, 0. , 0. , 0. , 0. ],
[-0.97596563, 2.04536053, 0.88367922, 1.013342 , -0.16605355,
3.02994344, 2.04080806, -0.25153046, -0.5964068 , 2.9607247 ,
-0.49722121, 0.02734492, 2.16949987, 2.77367066, 0.15628842,
2.19823207, 0. , 0. , 0. , 0. ],
[ 0.31546283, 3.27420503, 3.23550769, -0.63724013, 0.89150128,
0.69774266, 2.76627308, -0.58408384, -0.45681779, 1.98843041,
-0.31850477, 0.83729882, 0.45471165, 3.61974147, -1.45610756,
1.35217453, 0. , 0. , 0. , 0. ],
[ 1.03329532, 1.97471646, 1.33949611, 1.22857243, -1.46890642,
1.74105506, 1.40969261, 0.52465603, -0.18895266, 2.81025597,
2.64901037, -0.83415186, 0.76956826, 1.48730868, -0.16190164,
2.24389007, 0. , 0. , 0. , 0. ],
[-1.0676654 , 3.08429323, 1.7601179 , 0.85448051, 1.15537064,
2.82487842, 0.27891413, 0.57842569, -0.62392063, 1.00343057,
1.15348843, -0.37650332, 3.37355345, 2.22285473, 0.43444434,
0.15743873, 0. , 0. , 0. , 0. ]]),
array([[ 1.05258873, -0.17897376, -0.99932932, -1.02854121, 0.85159208,
2.32349131, 1.96526709, -0.08398597, -0.69474809, 1.32820222,
1.19514151, 1.56814867, 0.86013263, 1.48342922, 0. ,
0. , 0. , 0. , 0. , 0. ],
[ 0.1920635 , -0.48702788, 1.24353985, -1.3864121 , 0.16713229,
3.10134683, 0.61658271, -0.63360643, 0.86000807, 2.74876157,
2.87604877, 0.16339724, 2.87595396, 3.2846962 , 0. ,
0. , 0. , 0. , 0. , 0. ],
[ 0.1380241 , -0.76783029, 0.18814436, -1.18165209, -0.02981728,
1.49908113, 0.61521007, -0.98191097, 0.31250199, 1.39015803,
3.16213211, -0.70891214, 3.83881766, 1.92683533, 0. ,
0. , 0. , 0. , 0. , 0. ],
[ 1.39080778, -0.59179216, 0.80348201, 0.64638205, -1.40144268,
1.49751413, 3.0092166 , 1.33099666, 1.43714841, 2.90734268,
3.09688943, 0.32934884, 1.14592787, 1.58152023, 0. ,
0. , 0. , 0. , 0. , 0. ],
[-0.77164353, 0.50293096, 0.0717377 , 0.14487556, -0.90246591,
2.32612179, 1.98628857, 1.29683166, -0.12399569, 2.60184685,
3.20136653, 0.44056647, 0.98283455, 1.79026663, 0. ,
0. , 0. , 0. , 0. , 0. ]]),
array([[-0.93359914, 2.31840281, 0.55691601, 1.90930758, -1.58260431,
-1.05801881, 3.28012523, 3.84105406, -1.2127093 , 0.00490079,
1.28149304, 0. , 0. , 0. , 0. ,
0. , 0. , 0. , 0. , 0. ],
[-1.03105486, 2.7703693 , 0.16751813, 1.12127987, -0.44070271,
-0.0789227 , 2.79008301, 1.11456745, 1.13982551, -1.10128658,
0.87430834, 0. , 0. , 0. , 0. ,
0. , 0. , 0. , 0. , 0. ],
[-0.69710668, 1.72702833, -2.62599502, 2.34730002, 0.77756661,
0.16415884, 3.30712178, 1.67331828, -0.44022431, 0.56837829,
1.1566811 , 0. , 0. , 0. , 0. ,
0. , 0. , 0. , 0. , 0. ],
[-0.71845983, 1.79908544, 0.37385522, 1.3870915 , -1.48823234,
-1.487419 , 3.0879945 , 1.74617784, -0.91538815, -0.24244522,
0.81393954, 0. , 0. , 0. , 0. ,
0. , 0. , 0. , 0. , 0. ],
[-1.38501563, 3.73330047, -0.52494265, 2.37133716, -0.24546709,
-0.28360782, 2.89384717, 2.42891743, 0.40144022, -1.21850571,
2.00370751, 0. , 0. , 0. , 0. ,
0. , 0. , 0. , 0. , 0. ]]),
array([[ 1.27989188, 1.16254538, -0.06889142, 1.84133355, 1.3234908 ,
1.29611702, 2.0019294 , -0.03220116, 1.1085194 , 1.96495985,
1.68544302, 1.94503544, 0. , 0. , 0. ,
0. , 0. , 0. , 0. , 0. ],
[ 1.3004439 , 2.48768923, 0.59809607, 2.38155155, 2.78705889,
1.67018683, 0.21731778, -0.59277191, 2.87427207, 2.63950475,
2.39211459, 0.93083423, 0. , 0. , 0. ,
0. , 0. , 0. , 0. , 0. ],
[ 2.39239371, 0.30900383, -0.97307155, 1.98100711, 0.30613735,
1.12827171, 0.16987791, 0.31959096, 1.30366416, 1.45881023,
2.45668401, 0.5218711 , 0. , 0. , 0. ,
0. , 0. , 0. , 0. , 0. ],
[ 0.0826574 , 2.05100254, 0.013161 , 2.95120798, 1.15730011,
0.75537024, 0.13708569, -0.44922143, 0.64834001, 2.50640862,
2.00349347, 3.35573624, 0. , 0. , 0. ,
0. , 0. , 0. , 0. , 0. ],
[ 0.47135124, 2.10258532, 0.70212032, 2.56063126, 1.62466971,
2.64026892, 0.21309489, -0.57752813, 2.21335957, 0.20453233,
0.03106993, 3.01167822, 0. , 0. , 0. ,
0. , 0. , 0. , 0. , 0. ]]),
array([[-0.42125521, 0.54016939, 1.63016057, 2.01555253, -0.10961255,
-0.42549555, 1.55793753, -0.0998756 , 0.36417335, 3.37126414,
1.62151191, 2.84084192, 0.10831384, 0.89293054, -0.08671363,
0.49340353, 0. , 0. , 0. , 0. ],
[-0.37615411, 2.00581062, 2.30426605, 2.02205839, 0.65871664,
1.34478836, -0.55379752, -1.42787727, 0.59732227, 0.84969282,
0.54345723, 0.95849568, -0.17131602, -0.70425277, -0.5337757 ,
1.78207229, 0. , 0. , 0. , 0. ],
[-0.13863276, 1.71490034, 2.02677925, 2.60608619, 0.26916522,
0.35928298, -1.26521844, -0.59859219, 1.19162219, 1.64565259,
1.16787165, 2.95245196, 0.48681084, 1.66621053, 0.918077 ,
-1.10583747, 0. , 0. , 0. , 0. ],
[ 0.87763797, 2.38740754, 2.9111822 , 2.21184069, 0.78091173,
-0.53270909, 0.40100338, -0.83375593, 0.9860009 , 2.43898437,
-0.64499989, 2.95092003, -1.52360727, 0.44640918, 0.78131922,
-0.24401283, 0. , 0. , 0. , 0. ],
[ 0.92615066, 3.45437746, 3.28808981, 2.87207404, -1.60027223,
-1.14164941, -1.63807699, 0.33084805, 2.92963629, 3.51170824,
-0.3286093 , 2.19108385, 0.97812366, -1.82565766, -0.34034678,
-2.0485913 , 0. , 0. , 0. , 0. ]]),
array([[ 1.96438618e+00, 1.88104784e-01, 1.61114494e+00,
6.99567690e-04, 2.55271963e+00, 0.00000000e+00,
0.00000000e+00, 0.00000000e+00, 0.00000000e+00,
0.00000000e+00, 0.00000000e+00, 0.00000000e+00,
0.00000000e+00, 0.00000000e+00, 0.00000000e+00,
0.00000000e+00, 0.00000000e+00, 0.00000000e+00,
0.00000000e+00, 0.00000000e+00],
[ 2.41578815e+00, -5.70625661e-01, 2.15545894e+00,
-1.80948908e+00, 1.62049331e+00, 0.00000000e+00,
0.00000000e+00, 0.00000000e+00, 0.00000000e+00,
0.00000000e+00, 0.00000000e+00, 0.00000000e+00,
0.00000000e+00, 0.00000000e+00, 0.00000000e+00,
0.00000000e+00, 0.00000000e+00, 0.00000000e+00,
0.00000000e+00, 0.00000000e+00],
[ 1.97017040e+00, -1.62556528e+00, 2.49469152e+00,
4.18785985e-02, 2.61875866e+00, 0.00000000e+00,
0.00000000e+00, 0.00000000e+00, 0.00000000e+00,
0.00000000e+00, 0.00000000e+00, 0.00000000e+00,
0.00000000e+00, 0.00000000e+00, 0.00000000e+00,
0.00000000e+00, 0.00000000e+00, 0.00000000e+00,
0.00000000e+00, 0.00000000e+00],
[ 3.14277819e+00, 3.01098398e-02, 7.40376369e-01,
1.76517344e+00, 2.68922918e+00, 0.00000000e+00,
0.00000000e+00, 0.00000000e+00, 0.00000000e+00,
0.00000000e+00, 0.00000000e+00, 0.00000000e+00,
0.00000000e+00, 0.00000000e+00, 0.00000000e+00,
0.00000000e+00, 0.00000000e+00, 0.00000000e+00,
0.00000000e+00, 0.00000000e+00],
[ 2.06250296e+00, 4.67605528e-01, 1.55927230e+00,
1.85788889e-01, 1.30359922e+00, 0.00000000e+00,
0.00000000e+00, 0.00000000e+00, 0.00000000e+00,
0.00000000e+00, 0.00000000e+00, 0.00000000e+00,
0.00000000e+00, 0.00000000e+00, 0.00000000e+00,
0.00000000e+00, 0.00000000e+00, 0.00000000e+00,
0.00000000e+00, 0.00000000e+00]]),
array([[ 1.22152427, 3.74926839, 0.64415552, 2.35268329, 1.98754653,
2.89384829, 0.44589817, 3.94228743, 2.72405657, 0.86222004,
0.68681903, 3.89952458, 1.43454512, 0. , 0. ,
0. , 0. , 0. , 0. , 0. ],
[-0.02203262, 0.95065123, 0.71669023, 0.02919391, 2.30714524,
1.91843002, 0.73611294, 1.20560482, 0.85206836, -0.74221506,
-0.72886308, 2.39872927, -0.95841402, 0. , 0. ,
0. , 0. , 0. , 0. , 0. ],
[ 0.55775319, 0.33773314, 0.79932151, 1.94966883, 3.2113281 ,
2.70768249, -0.69745554, 1.23208345, 1.66199957, 1.69894081,
0.13124461, 1.93256147, -0.17787952, 0. , 0. ,
0. , 0. , 0. , 0. , 0. ],
[ 0.45089205, 2.62430534, -1.9517961 , 2.24040577, 1.75642049,
1.94962325, 0.26796497, 2.28418304, 1.44944487, 0.28723885,
-0.81081633, 1.54840214, 0.82652939, 0. , 0. ,
0. , 0. , 0. , 0. , 0. ],
[ 1.27678173, 1.17204606, -0.24738322, 1.02761617, 1.81060444,
2.37830861, 0.55260134, 2.50046334, 1.04652821, 0.03467176,
-2.07336654, 1.2628897 , 0.61604732, 0. , 0. ,
0. , 0. , 0. , 0. , 0. ]]),
array([[ 3.86138405, 2.35068317, -1.90187438, 0.600788 , 0.18011722,
1.3469559 , -0.54708828, 1.83798823, -0.01957845, 2.88713217,
3.1724991 , 2.90802072, 0. , 0. , 0. ,
0. , 0. , 0. , 0. , 0. ],
[ 1.26785642, 0.51076756, 0.32070756, 2.33758816, 2.08146669,
-0.60796736, 0.93777509, 2.70474711, 0.44785738, 1.61720609,
1.52890594, 3.03072971, 0. , 0. , 0. ,
0. , 0. , 0. , 0. , 0. ],
[ 3.30219394, 3.1515445 , 1.16550716, 2.07489374, 0.66441859,
0.97529244, 0.35176367, 1.22593639, -1.80698271, 1.19936482,
3.34017172, 2.15960657, 0. , 0. , 0. ,
0. , 0. , 0. , 0. , 0. ],
[ 2.34839018, 2.24827352, -1.61070856, 2.81044265, -1.21423372,
0.24633846, -0.82196609, 2.28616568, 0.033922 , 2.7557593 ,
1.16178372, 3.66959512, 0. , 0. , 0. ,
0. , 0. , 0. , 0. , 0. ],
[ 1.32913219, 1.63231852, 0.58642744, 1.55873546, 0.86354741,
2.06654246, -0.44036504, 3.22723595, 1.33279468, 0.05975892,
2.48518999, 3.44690602, 0. , 0. , 0. ,
0. , 0. , 0. , 0. , 0. ]]),
array([[ 0.61424344, -1.03068819, -1.47929328, 2.91514641, 2.06867196,
1.90384921, -0.45835234, 1.22054782, 0.67931536, 0. ,
0. , 0. , 0. , 0. , 0. ,
0. , 0. , 0. , 0. , 0. ],
[ 2.76480464, 1.12442631, -2.36004758, 2.91912726, 1.67891181,
3.76873596, -0.93874096, -0.32397781, -0.55732374, 0. ,
0. , 0. , 0. , 0. , 0. ,
0. , 0. , 0. , 0. , 0. ],
[ 0.39953353, -1.26828104, 0.44482517, 2.85604975, 3.08891062,
2.60268725, -0.15785176, 1.58549879, -0.32948578, 0. ,
0. , 0. , 0. , 0. , 0. ,
0. , 0. , 0. , 0. , 0. ],
[ 1.65156484, -1.56545168, -1.42771206, 2.74216475, 1.8758154 ,
3.51169147, 0.18353058, -0.14704149, 0.00442783, 0. ,
0. , 0. , 0. , 0. , 0. ,
0. , 0. , 0. , 0. , 0. ],
[ 1.27736372, 0.37407608, -1.25713475, 0.53171176, 1.53714914,
0.21015523, -1.06850669, -0.09755327, -0.92373834, 0. ,
0. , 0. , 0. , 0. , 0. ,
0. , 0. , 0. , 0. , 0. ]]),
array([[-1.39160433, 0.21014669, -0.89792475, 2.6702794 , 1.54610601,
0.84699037, 2.96726482, 1.84236946, 0.02211578, 0.32842575,
1.02718924, 1.78447936, -1.20056829, 2.26699318, -0.23156537,
2.50124959, 1.93372501, 0.10264369, -1.70813962, 0. ],
[ 0.38823591, -1.30348049, -0.31599117, 2.60044143, 2.32929389,
1.40348483, 3.25758736, 1.92210728, -0.34150988, -1.22336921,
2.3567069 , 1.75456835, 0.28295694, 0.68114898, -0.457843 ,
1.83372069, 2.10177851, -0.26664178, -0.26549595, 0. ],
[ 0.08540346, 0.71507504, 1.78164285, 3.04418137, 1.52975256,
3.55159169, 3.21396003, 3.22720346, 0.68147142, 0.12466013,
-0.4122895 , 1.97986653, 1.51671949, 2.06096825, -0.6765908 ,
2.00145086, 1.73723014, 0.50186043, -2.27525744, 0. ],
[ 0.00632717, 0.3050794 , -0.33167875, 1.48109172, 0.19653696,
1.97504239, 2.51595821, 1.74499313, -1.65198805, -1.04424953,
-0.23786945, 1.18639347, -0.03568057, 3.82541131, 2.84039446,
2.88325909, 1.79827675, -0.80230291, 0.08165052, 0. ],
[ 0.89980086, 0.34690991, -0.60806566, 1.69472308, 1.38043417,
0.97139487, 0.21977176, 1.01340944, -1.69946943, -0.01775586,
-0.35851919, 1.81115864, 1.15105661, 1.21410373, 1.50667558,
1.70155313, 3.1410754 , -0.54806167, -0.51879299, 0. ]])]
y_batches_train
[array([1., 2., 2., 1., 1., 2., 2., 1., 1., 2., 1., 1., 2., 2., 1., 2., 0.,
0., 0., 0.]),
array([1., 1., 1., 1., 1., 2., 2., 1., 1., 2., 2., 1., 2., 2., 0., 0., 0.,
0., 0., 0.]),
array([1., 2., 1., 2., 1., 1., 2., 2., 1., 1., 2., 0., 0., 0., 0., 0., 0.,
0., 0., 0.]),
array([2., 2., 1., 2., 2., 2., 1., 1., 2., 2., 2., 2., 0., 0., 0., 0., 0.,
0., 0., 0.]),
array([1., 2., 2., 2., 1., 1., 1., 1., 2., 2., 1., 2., 1., 1., 1., 1., 0.,
0., 0., 0.]),
array([2., 1., 2., 1., 2., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,
0., 0., 0.]),
array([1., 2., 1., 2., 2., 2., 1., 2., 2., 1., 1., 2., 1., 0., 0., 0., 0.,
0., 0., 0.]),
array([2., 2., 1., 2., 1., 1., 1., 2., 1., 2., 2., 2., 0., 0., 0., 0., 0.,
0., 0., 0.]),
array([2., 1., 1., 2., 2., 2., 1., 1., 1., 0., 0., 0., 0., 0., 0., 0., 0.,
0., 0., 0.]),
array([1., 1., 1., 2., 2., 2., 2., 2., 1., 1., 1., 2., 1., 2., 1., 2., 2.,
1., 1., 0.])]
X_batches_test
[array([[ 0.74119496, 1.97273418, 1.76675805, 0.51484268, 1.39422086,
2.97184667, -1.35274514, 2.08825434, -1.2521965 , 1.11556387,
0.19776789, 2.38259223, -0.57140597, -0.79010112, 0.17038974,
1.28075761, 0.696398 , 3.0920007 , -0.41138503, 0. ],
[-1.39081797, 0.41079718, 3.03698894, -2.07333633, 2.05575621,
2.73222939, -0.98182787, 1.06741172, -1.36310914, 0.20174856,
0.35323654, 2.70305775, 0.52549713, -0.7786237 , 1.80857093,
0.96830907, -0.23610863, 1.28160768, 0.7026651 , 0. ],
[ 1.16357113, 0.43907935, 3.40158623, -0.73923043, 1.484668 ,
1.52809569, -0.02347205, 1.65349967, 1.79635118, -0.46647772,
-0.78400883, 0.82695404, -1.34932627, -0.3200281 , 2.84417045,
0.01534261, 0.10047148, 2.70769609, -1.42669461, 0. ],
[-1.05475682, 3.45578027, 1.58589338, -0.55515227, 2.13477478,
1.86777473, 0.61550335, 1.05781415, -0.45297406, -0.04317595,
-0.15255388, 0.74669395, -1.43621979, 1.06229278, 0.99792794,
1.24391783, -1.86484584, 1.92802343, 0.56148011, 0. ],
[-0.0835337 , 1.89593955, 1.65769335, -0.93622246, 1.05002869,
1.49675624, -0.00821712, 1.71541053, 2.02408452, 0.59011484,
0.72719784, 3.44801858, -0.00957537, 0.37176007, 1.93481168,
2.23125062, 1.67910471, 2.80923862, 0.34516993, 0. ]]),
array([[ 0.40691415, 2.31873444, -0.83458005, -0.17018249, -0.39177831,
1.90353251, 2.98241467, 0.32808584, 3.09429553, 2.27183083,
3.09576659, 0. , 0. , 0. , 0. ,
0. , 0. , 0. , 0. , 0. ],
[ 1.6862473 , 1.0690102 , -0.07415598, -0.09846767, 1.14562424,
2.52211963, 1.71911351, 0.41879894, 1.62787544, 3.50533394,
2.69963456, 0. , 0. , 0. , 0. ,
0. , 0. , 0. , 0. , 0. ],
[ 3.27824216, 2.25067953, 0.40017321, -1.36011162, -1.41010106,
0.98956203, 2.30881584, -0.29496046, 2.29748247, 3.24940966,
1.06431776, 0. , 0. , 0. , 0. ,
0. , 0. , 0. , 0. , 0. ],
[ 2.80167214, 3.88324559, -0.6984172 , 0.81889567, 1.86945352,
3.07554419, 3.10357189, 1.31426767, 0.28163147, 2.75559628,
2.00866885, 0. , 0. , 0. , 0. ,
0. , 0. , 0. , 0. , 0. ],
[ 1.54574419, 1.00720596, -1.55418837, 0.70823839, 0.14715209,
1.03747262, 0.82988672, -0.54006372, 1.4960777 , 0.34578788,
1.10558132, 0. , 0. , 0. , 0. ,
0. , 0. , 0. , 0. , 0. ]])]
y_batches_test
[array([1., 2., 2., 1., 2., 2., 1., 2., 1., 1., 1., 2., 1., 1., 2., 2., 1.,
2., 1., 0.]),
array([2., 2., 1., 1., 1., 2., 2., 1., 2., 2., 2., 0., 0., 0., 0., 0., 0.,
0., 0., 0.])]
696
Masking is a way to tell sequence-processing layers that certain timesteps in an input are missing, and thus should be skipped when processing the data.
Batch Size is the number of samples we send to the model at a time. In this example, we have batch size = 2 but you can take it 4, 8,16, 32, 64 etc depends on the memory (basically in 2's power)
To ensure that all the input sequence data is having the same length we pad or truncate the input data points. The deep learning model accepts the input data points of standardized tensors. We define the tensors as: tensor(batch_size, length_sequence,features)
mask_zero: Boolean, whether or not the input value 0 is a special "padding" value that should be masked out. This is useful when using recurrent layers which may take variable length input. If this is True , then all subsequent layers in the model need to support masking or an exception will be raised.
First problem: your X data after reshaping is not what you expected. If you look at the first sample after reshaping, it is:
array([[[-1.00612917, 1.47313952, 2.68021318, 1.54875809,
0.98385996],
[ 1.49465265, 0.60429106, 1.12396908, -0.24041602,
1.77266187],
[ 0.1961381 , 1.28019637, 1.78803092, 2.05151245,
0.93606708],
[ 0.51554755, 0. , 0. , 0. ,
0. ],
[-0.97596563, 2.04536053, 0.88367922, 1.013342 ,
-0.16605355],
[ 3.02994344, 2.04080806, -0.25153046, -0.5964068 ,
2.9607247 ],
[-0.49722121, 0.02734492, 2.16949987, 2.77367066,
0.15628842],
[ 2.19823207, 0. , 0. , 0. ,
0. ],
[ 0.31546283, 3.27420503, 3.23550769, -0.63724013,
0.89150128],
[ 0.69774266, 2.76627308, -0.58408384, -0.45681779,
1.98843041],
[-0.31850477, 0.83729882, 0.45471165, 3.61974147,
-1.45610756],
[ 1.35217453, 0. , 0. , 0. ,
0. ],
[ 1.03329532, 1.97471646, 1.33949611, 1.22857243,
-1.46890642],
[ 1.74105506, 1.40969261, 0.52465603, -0.18895266,
2.81025597],
[ 2.64901037, -0.83415186, 0.76956826, 1.48730868,
-0.16190164],
[ 2.24389007, 0. , 0. , 0. ,
0. ],
[-1.0676654 , 3.08429323, 1.7601179 , 0.85448051,
1.15537064],
[ 2.82487842, 0.27891413, 0.57842569, -0.62392063,
1.00343057],
[ 1.15348843, -0.37650332, 3.37355345, 2.22285473,
0.43444434],
[ 0.15743873, 0. , 0. , 0. ,
0. ]]])
So actually no timestep is masked, because Masking layer only mask timesteps where all features are 0, so the above 20 timesteps are not masked because none of them are completely 0.
For the Masking layer, to ensure you have the mask propogated to the output layer successfully, you can do:
for i, l in enumerate(model.layers):
print(f'layer {i}: {l}')
print(f'has input mask: {l.input_mask}')
print(f'has output mask: {l.output_mask}')
layer 0: <tensorflow.python.keras.layers.core.Masking object at 0x6417b7f60>
has input mask: None
has output mask: Tensor("masking/Identity_1:0", shape=(None, 20), dtype=bool)
layer 1: <tensorflow.python.keras.layers.wrappers.Bidirectional object at 0x641e25cf8>
has input mask: Tensor("masking/Identity_1:0", shape=(None, 20), dtype=bool)
has output mask: Tensor("masking/Identity_1:0", shape=(None, 20), dtype=bool)
layer 2: <tensorflow.python.keras.layers.core.Dropout object at 0x641814128>
has input mask: Tensor("masking/Identity_1:0", shape=(None, 20), dtype=bool)
has output mask: Tensor("masking/Identity_1:0", shape=(None, 20), dtype=bool)
layer 3: <tensorflow.python.keras.layers.wrappers.TimeDistributed object at 0x6433b6ba8>
has input mask: Tensor("masking/Identity_1:0", shape=(None, 20), dtype=bool)
has output mask: Tensor("time_distributed/Reshape_3:0", shape=(None, 20), dtype=bool)
So you can see that the final layer also has the output_mask, which means the masks are successfully propogated. You seem to have a misunderstanding of how Masking works in Keras, what it actually does is it will generate a mask, which is a boolean array, the shape of the mask is (None, Timesteps), since in your model definition, the Timestep dimension is always kept the same, so the mask will be propogated to the end without any changes. Then when Keras calculate loss (and of course when it calculate gradients), the timesteps which has a mask value False will be ignored. The Masking layer doesn't change the output value and of course your model will still predict class 0, what it only does is to produce a boolean array indicating which timestep should be skipped and pass it to the end (if all the layers accept the mask).
So what you can do is change one line of your model definition as follows and make your y_labels shifted by 1, which means your current classes:
0 -> 0 (since the loss of these timesteps will be ignored, not contributing to the training of the model, so whether it is 0 or 1 doesn't matter)
1 -> 0
2 -> 1
# I would prefer softmax if doing classification
# here we only need to specify 2 classes
# and actually TimeDistributed can be thrown away (at least in recent Keras versions)
model.add(TimeDistributed(Dense(2, activation='softmax')))
You can also see my answer here https://stackoverflow.com/a/59313862/11819266 for understanding how the loss are calculated with / without masking.
74
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With
