ValueError: Can not squeeze dim[1], expected a dimension of 1, got 3 for 'sparse_softmax_cross_entropy_loss
Question:
I tried to replace the training and validation data with local images. But when running the training code, it came up with the error :
ValueError: Can not squeeze dim[1], expected a dimension of 1, got 3 for ‘sparse_softmax_cross_entropy_loss/remove_squeezable_dimensions/Squeeze’ (op: ‘Squeeze’) with input shapes: [100,3].
I don’t know how to fix it up. There is no visible variable in the model definition code. The code was modified from TensorFlow tutorial. The images are jpgs.
Here is the detail Error message:
INFO:tensorflow:Using default config.
INFO:tensorflow:Using config: {'_log_step_count_steps': 100, '_is_chief': True, '_model_dir': '/tmp/mnist_convnet_model', '_tf_random_seed': None, '_session_config': None, '_save_checkpoints_secs': 600, '_num_worker_replicas': 1, '_save_checkpoints_steps': None, '_service': None, '_keep_checkpoint_max': 5, '_cluster_spec': <tensorflow.python.training.server_lib.ClusterSpec object at 0x00000288088D50F0>, '_keep_checkpoint_every_n_hours': 10000, '_task_type': 'worker', '_master': '', '_save_summary_steps': 100, '_num_ps_replicas': 0, '_task_id': 0}
Traceback (most recent call last):
File "C:UsersASUSAppDataLocalProgramsPythonPython35libsite-packagestensorflowpythonframeworkcommon_shapes.py", line 686, in _call_cpp_shape_fn_impl
input_tensors_as_shapes, status)
File "C:UsersASUSAppDataLocalProgramsPythonPython35libsite-packagestensorflowpythonframeworkerrors_impl.py", line 473, in __exit__
c_api.TF_GetCode(self.status.status))
tensorflow.python.framework.errors_impl.InvalidArgumentError: Can not squeeze dim[1], expected a dimension of 1, got 3 for 'sparse_softmax_cross_entropy_loss/remove_squeezable_dimensions/Squeeze' (op: 'Squeeze') with input shapes: [100,3].
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "D:tf_exe_5_make_image_lablescnn_mnist.py", line 214, in <module>
tf.app.run()
File "C:UsersASUSAppDataLocalProgramsPythonPython35libsite-packagestensorflowpythonplatformapp.py", line 124, in run
_sys.exit(main(argv))
File "D:tf_exe_5_make_image_lablescnn_mnist.py", line 203, in main
hooks=[logging_hook])
File "C:UsersASUSAppDataLocalProgramsPythonPython35libsite-packagestensorflowpythonestimatorestimator.py", line 314, in train
loss = self._train_model(input_fn, hooks, saving_listeners)
File "C:UsersASUSAppDataLocalProgramsPythonPython35libsite-packagestensorflowpythonestimatorestimator.py", line 743, in _train_model
features, labels, model_fn_lib.ModeKeys.TRAIN, self.config)
File "C:UsersASUSAppDataLocalProgramsPythonPython35libsite-packagestensorflowpythonestimatorestimator.py", line 725, in _call_model_fn
model_fn_results = self._model_fn(features=features, **kwargs)
File "D:tf_exe_5_make_image_lablescnn_mnist.py", line 67, in cnn_model_fn
loss = tf.losses.sparse_softmax_cross_entropy(labels=labels, logits=logits)
File "C:UsersASUSAppDataLocalProgramsPythonPython35libsite-packagestensorflowpythonopslosseslosses_impl.py", line 790, in sparse_softmax_cross_entropy
labels, logits, weights, expected_rank_diff=1)
File "C:UsersASUSAppDataLocalProgramsPythonPython35libsite-packagestensorflowpythonopslosseslosses_impl.py", line 720, in _remove_squeezable_dimensions
labels, predictions, expected_rank_diff=expected_rank_diff)
File "C:UsersASUSAppDataLocalProgramsPythonPython35libsite-packagestensorflowpythonopsconfusion_matrix.py", line 76, in remove_squeezable_dimensions
labels = array_ops.squeeze(labels, [-1])
File "C:UsersASUSAppDataLocalProgramsPythonPython35libsite-packagestensorflowpythonopsarray_ops.py", line 2490, in squeeze
return gen_array_ops._squeeze(input, axis, name)
File "C:UsersASUSAppDataLocalProgramsPythonPython35libsite-packagestensorflowpythonopsgen_array_ops.py", line 7049, in _squeeze
"Squeeze", input=input, squeeze_dims=axis, name=name)
File "C:UsersASUSAppDataLocalProgramsPythonPython35libsite-packagestensorflowpythonframeworkop_def_library.py", line 787, in _apply_op_helper
op_def=op_def)
File "C:UsersASUSAppDataLocalProgramsPythonPython35libsite-packagestensorflowpythonframeworkops.py", line 3162, in create_op
compute_device=compute_device)
File "C:UsersASUSAppDataLocalProgramsPythonPython35libsite-packagestensorflowpythonframeworkops.py", line 3208, in _create_op_helper
set_shapes_for_outputs(op)
File "C:UsersASUSAppDataLocalProgramsPythonPython35libsite-packagestensorflowpythonframeworkops.py", line 2427, in set_shapes_for_outputs
return _set_shapes_for_outputs(op)
File "C:UsersASUSAppDataLocalProgramsPythonPython35libsite-packagestensorflowpythonframeworkops.py", line 2400, in _set_shapes_for_outputs
shapes = shape_func(op)
File "C:UsersASUSAppDataLocalProgramsPythonPython35libsite-packagestensorflowpythonframeworkops.py", line 2330, in call_with_requiring
return call_cpp_shape_fn(op, require_shape_fn=True)
File "C:UsersASUSAppDataLocalProgramsPythonPython35libsite-packagestensorflowpythonframeworkcommon_shapes.py", line 627, in call_cpp_shape_fn
require_shape_fn)
File "C:UsersASUSAppDataLocalProgramsPythonPython35libsite-packagestensorflowpythonframeworkcommon_shapes.py", line 691, in _call_cpp_shape_fn_impl
raise ValueError(err.message)
ValueError: Can not squeeze dim[1], expected a dimension of 1, got 3 for 'sparse_softmax_cross_entropy_loss/remove_squeezable_dimensions/Squeeze' (op: 'Squeeze') with input shapes: [100,3].
>>>
Here is my code:
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
#imports
import numpy as np
import tensorflow as tf
import glob
import cv2
import random
import matplotlib.pylab as plt
import pandas as pd
import sys as system
from mlxtend.preprocessing import one_hot
from sklearn import preprocessing
from sklearn.preprocessing import LabelEncoder
from sklearn.preprocessing import OneHotEncoder
tf.logging.set_verbosity(tf.logging.INFO)
def cnn_model_fn(features, labels, mode):
"""Model function for CNN"""
#Input Layer
input_layer = tf.reshape(features["x"], [-1,320,320,3])
#Convolutional Layer #1
conv1 = tf.layers.conv2d(
inputs = input_layer,
filters = 32,
kernel_size=[5,5],
padding = "same",
activation=tf.nn.relu)
#Pooling Layer #1
pool1 = tf.layers.max_pooling2d(inputs=conv1, pool_size=[2,2], strides=2)
#Convolutional Layer #2 and Pooling Layer #2
conv2 = tf.layers.conv2d(
inputs=pool1,
filters=64,
kernel_size=[5,5],
padding="same",
activation=tf.nn.relu)
pool2 = tf.layers.max_pooling2d(inputs=conv2, pool_size=[2,2], strides=2)
#Dense Layer
pool2_flat = tf.reshape(pool2, [-1,80*80*64])
dense = tf.layers.dense(inputs=pool2_flat, units=1024, activation=tf.nn.relu)
dropout = tf.layers.dropout(
inputs=dense, rate=0.4, training=mode == tf.estimator.ModeKeys.TRAIN)
#Logits Layer
logits = tf.layers.dense(inputs=dropout, units=3)
predictions = {
#Generate predictions (for PREDICT and EVAL mode)
"classes": tf.argmax(input=logits, axis=1),
#Add 'softmax_tensor' to the graph. It is used for PREDICT and by the
#'logging_hook'
"probabilities": tf.nn.softmax(logits, name="softmax_tensor")
}
if mode == tf.estimator.ModeKeys.PREDICT:
return tf.estimator.EstimatorSpec(mode=mode, predictions=predictions)
# Calculate Loss (for both TRAIN and EVAL modes
loss = tf.losses.sparse_softmax_cross_entropy(labels=labels, logits=logits)
# Configure the Training Op (for TRAIN mode)
if mode == tf.estimator.ModeKeys.TRAIN:
optimizer = tf.train.GradientDescentOptimizer(learning_rate=0.001)
train_op = optimizer.minimize(
loss=loss,
global_step=tf.train.get_global_step())
return tf.estimator.EstimatorSpec(mode=mode, loss=loss, train_op=train_op)
# Add evaluation metrics (for EVAL mode)
eval_metric_ops = {
"accuracy": tf.metrics.accuracy(
labels=labels, predictions=predictions["classes"])}
return tf.estimator.EstimatorSpec(
mode=mode, loss=loss,eval_metric_ops=eval_metric_ops)
def main(unused_argv):
'''
#Load training and eval data
mnist = tf.contrib.learn.datasets.load_dataset("mnist")
train_data = mnist.train.images
train_labels = np.asarray(mnist.train.labels, dtype=np.int32)
eval_data = mnist.test.images
eval_labels = np.asarray(mnist.test.labels, dtype=np.int32)
'''
#Load cats, dogs and cars image in local folder
X_data = []
files = glob.glob("data/cats/*.jpg")
for myFile in files:
image = cv2.imread (myFile)
imgR = cv2.resize(image, (320, 320))
imgNR = imgR/255
X_data.append(imgNR)
files = glob.glob("data/dogs/*.jpg")
for myFile in files:
image = cv2.imread (myFile)
imgR = cv2.resize(image, (320, 320))
imgNR = imgR/255
X_data.append(imgNR)
files = glob.glob ("data/cars/*.jpg")
for myFile in files:
image = cv2.imread (myFile)
imgR = cv2.resize(image, (320, 320))
imgNR = imgR/255
X_data.append (imgNR)
#print('X_data count:', len(X_data))
X_data_Val = []
files = glob.glob ("data/Validation/cats/*.jpg")
for myFile in files:
image = cv2.imread (myFile)
imgR = cv2.resize(image, (320, 320))
imgNR = imgR/255
X_data_Val.append (imgNR)
files = glob.glob ("data/Validation/dogs/*.jpg")
for myFile in files:
image = cv2.imread (myFile)
imgR = cv2.resize(image, (320, 320))
imgNR = imgR/255
X_data_Val.append (imgNR)
files = glob.glob ("data/Validation/cars/*.jpg")
for myFile in files:
image = cv2.imread (myFile)
imgR = cv2.resize(image, (320, 320))
imgNR = imgR/255
X_data_Val.append (imgNR)
#Feed One hot lables
Y_Label = np.zeros(shape=(300,1))
for el in range(0,100):
Y_Label[el]=[0]
for el in range(101,200):
Y_Label[el]=[1]
for el in range(201,300):
Y_Label[el]=[2]
onehot_encoder = OneHotEncoder(sparse=False)
#Y_Label_RS = Y_Label.reshape(len(Y_Label), 1)
Y_Label_Encode = onehot_encoder.fit_transform(Y_Label)
#print('Y_Label_Encode shape:', Y_Label_Encode.shape)
Y_Label_Val = np.zeros(shape=(30,1))
for el in range(0, 10):
Y_Label_Val[el]=[0]
for el in range(11, 20):
Y_Label_Val[el]=[1]
for el in range(21, 30):
Y_Label_Val[el]=[2]
#Y_Label_Val_RS = Y_Label_Val.reshape(len(Y_Label_Val), 1)
Y_Label_Val_Encode = onehot_encoder.fit_transform(Y_Label_Val)
#print('Y_Label_Val_Encode shape:', Y_Label_Val_Encode.shape)
train_data = np.array(X_data)
train_data = train_data.astype(np.float32)
train_labels = np.asarray(Y_Label_Encode, dtype=np.int32)
eval_data = np.array(X_data_Val)
eval_data = eval_data.astype(np.float32)
eval_labels = np.asarray(Y_Label_Val_Encode, dtype=np.int32)
print(train_data.shape)
print(train_labels.shape)
#Create the Estimator
mnist_classifier = tf.estimator.Estimator(
model_fn=cnn_model_fn, model_dir="/tmp/mnist_convnet_model")
# Set up logging for predictions
tensor_to_log = {"probabilities": "softmax_tensor"}
logging_hook = tf.train.LoggingTensorHook(
tensors=tensor_to_log, every_n_iter=50)
# Train the model
train_input_fn = tf.estimator.inputs.numpy_input_fn(
x={"x": train_data},
y=train_labels,
batch_size=100,
num_epochs=None,
shuffle=True)
mnist_classifier.train(
input_fn=train_input_fn,
#original steps are 20000
steps=1,
hooks=[logging_hook])
# Evaluate the model and print results
eval_input_fn = tf.estimator.inputs.numpy_input_fn(
x={"x": eval_data},
y=eval_labels,
num_epochs=1,
shuffle=False)
eval_results = mnist_classifier.evaluate(input_fn=eval_input_fn)
print(eval_results)
if __name__ == "__main__":
tf.app.run()
Answers:
The error here is from tf.losses.sparse_softmax_cross_entropy(labels=labels, logits=logits).
The TensorFlow documentation clearly states that “labels vector must provide a single specific index for the true class for each row of logits”. So your labels vector must include only class-indices like 0,1,2 and not their respective one-hot-encodings like [1,0,0], [0,1,0], [0,0,1].
Reproducing the error to explain further:
import numpy as np
import tensorflow as tf
# Create random-array and assign as logits tensor
np.random.seed(12345)
logits = tf.convert_to_tensor(np.random.sample((4,4)))
print logits.get_shape() #[4,4]
# Create random-labels (Assuming only 4 classes)
labels = tf.convert_to_tensor(np.array([2, 2, 0, 1]))
loss_1 = tf.losses.sparse_softmax_cross_entropy(labels, logits)
sess = tf.Session()
sess.run(tf.global_variables_initializer())
print 'Loss: {}'.format(sess.run(loss_1)) # 1.44836854
# Now giving one-hot-encodings in place of class-indices for labels
wrong_labels = tf.convert_to_tensor(np.array([[0,0,1,0], [0,0,1,0], [1,0,0,0],[0,1,0,0]]))
loss_2 = tf.losses.sparse_softmax_cross_entropy(wrong_labels, logits)
# This should give you a similar error as soon as you define it
So try giving class-indices instead of one-hot encodings in your Y_Labels vector.
Hope this clears your doubt.
I have solved this error. The labels were in onehot encoding, so it was in dimension of [,10], rather than [,1]. So I used tf.argmax().
If you used Keras’ ImageDataGenerator, you can add class_mode="sparse" to obtain the correct levels:
train_datagen = keras.preprocessing.image.ImageDataGenerator(
rescale=1./255,
shear_range=0.2,
zoom_range=0.2,
horizontal_flip=True)
train_generator = train_datagen.flow_from_directory(
'data/train',
target_size=(150, 150),
batch_size=32,
class_mode="sparse")
Alternatively, you might be able to use softmax_cross_entropy, which seems to use onehot encoding for the labels.
i write the code that change [1,0,0], [0,1,0], [0,0,1] to 0,1,2.
import numpy as np
import tensorflow as tf
def change_to_right(wrong_labels):
right_labels=[]
for x in wrong_labels:
for i in range(0,len(wrong_labels[0])):
if x[i]==1:
right_labels.append(i)
return right_labels
wrong_labels =np.array([[0,0,1,0], [0,0,1,0], [1,0,0,0],[0,1,0,0]])
right_labels =tf.convert_to_tensor(np.array(change_to_right(wrong_labels)))
Changing
loss='sparse_categorical_crossentropy'
to
loss='categorical_crossentropy'
worked for me.
In simple english, your loss function should be categorical_crossentropy if you have applied labelbinarizer (for hot encoding) to your test data. If you have not hot encoded your test data, you should use’sparse_categorical_crossentropy’.
You can change to loss=’categorical_crossentropy’ for one hot encoding or the other option as mentioned earlier is
tf.losses.sparse_softmax_cross_entropy(labels, logits),
I tried to replace the training and validation data with local images. But when running the training code, it came up with the error :
ValueError: Can not squeeze dim[1], expected a dimension of 1, got 3 for ‘sparse_softmax_cross_entropy_loss/remove_squeezable_dimensions/Squeeze’ (op: ‘Squeeze’) with input shapes: [100,3].
I don’t know how to fix it up. There is no visible variable in the model definition code. The code was modified from TensorFlow tutorial. The images are jpgs.
Here is the detail Error message:
INFO:tensorflow:Using default config.
INFO:tensorflow:Using config: {'_log_step_count_steps': 100, '_is_chief': True, '_model_dir': '/tmp/mnist_convnet_model', '_tf_random_seed': None, '_session_config': None, '_save_checkpoints_secs': 600, '_num_worker_replicas': 1, '_save_checkpoints_steps': None, '_service': None, '_keep_checkpoint_max': 5, '_cluster_spec': <tensorflow.python.training.server_lib.ClusterSpec object at 0x00000288088D50F0>, '_keep_checkpoint_every_n_hours': 10000, '_task_type': 'worker', '_master': '', '_save_summary_steps': 100, '_num_ps_replicas': 0, '_task_id': 0}
Traceback (most recent call last):
File "C:UsersASUSAppDataLocalProgramsPythonPython35libsite-packagestensorflowpythonframeworkcommon_shapes.py", line 686, in _call_cpp_shape_fn_impl
input_tensors_as_shapes, status)
File "C:UsersASUSAppDataLocalProgramsPythonPython35libsite-packagestensorflowpythonframeworkerrors_impl.py", line 473, in __exit__
c_api.TF_GetCode(self.status.status))
tensorflow.python.framework.errors_impl.InvalidArgumentError: Can not squeeze dim[1], expected a dimension of 1, got 3 for 'sparse_softmax_cross_entropy_loss/remove_squeezable_dimensions/Squeeze' (op: 'Squeeze') with input shapes: [100,3].
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "D:tf_exe_5_make_image_lablescnn_mnist.py", line 214, in <module>
tf.app.run()
File "C:UsersASUSAppDataLocalProgramsPythonPython35libsite-packagestensorflowpythonplatformapp.py", line 124, in run
_sys.exit(main(argv))
File "D:tf_exe_5_make_image_lablescnn_mnist.py", line 203, in main
hooks=[logging_hook])
File "C:UsersASUSAppDataLocalProgramsPythonPython35libsite-packagestensorflowpythonestimatorestimator.py", line 314, in train
loss = self._train_model(input_fn, hooks, saving_listeners)
File "C:UsersASUSAppDataLocalProgramsPythonPython35libsite-packagestensorflowpythonestimatorestimator.py", line 743, in _train_model
features, labels, model_fn_lib.ModeKeys.TRAIN, self.config)
File "C:UsersASUSAppDataLocalProgramsPythonPython35libsite-packagestensorflowpythonestimatorestimator.py", line 725, in _call_model_fn
model_fn_results = self._model_fn(features=features, **kwargs)
File "D:tf_exe_5_make_image_lablescnn_mnist.py", line 67, in cnn_model_fn
loss = tf.losses.sparse_softmax_cross_entropy(labels=labels, logits=logits)
File "C:UsersASUSAppDataLocalProgramsPythonPython35libsite-packagestensorflowpythonopslosseslosses_impl.py", line 790, in sparse_softmax_cross_entropy
labels, logits, weights, expected_rank_diff=1)
File "C:UsersASUSAppDataLocalProgramsPythonPython35libsite-packagestensorflowpythonopslosseslosses_impl.py", line 720, in _remove_squeezable_dimensions
labels, predictions, expected_rank_diff=expected_rank_diff)
File "C:UsersASUSAppDataLocalProgramsPythonPython35libsite-packagestensorflowpythonopsconfusion_matrix.py", line 76, in remove_squeezable_dimensions
labels = array_ops.squeeze(labels, [-1])
File "C:UsersASUSAppDataLocalProgramsPythonPython35libsite-packagestensorflowpythonopsarray_ops.py", line 2490, in squeeze
return gen_array_ops._squeeze(input, axis, name)
File "C:UsersASUSAppDataLocalProgramsPythonPython35libsite-packagestensorflowpythonopsgen_array_ops.py", line 7049, in _squeeze
"Squeeze", input=input, squeeze_dims=axis, name=name)
File "C:UsersASUSAppDataLocalProgramsPythonPython35libsite-packagestensorflowpythonframeworkop_def_library.py", line 787, in _apply_op_helper
op_def=op_def)
File "C:UsersASUSAppDataLocalProgramsPythonPython35libsite-packagestensorflowpythonframeworkops.py", line 3162, in create_op
compute_device=compute_device)
File "C:UsersASUSAppDataLocalProgramsPythonPython35libsite-packagestensorflowpythonframeworkops.py", line 3208, in _create_op_helper
set_shapes_for_outputs(op)
File "C:UsersASUSAppDataLocalProgramsPythonPython35libsite-packagestensorflowpythonframeworkops.py", line 2427, in set_shapes_for_outputs
return _set_shapes_for_outputs(op)
File "C:UsersASUSAppDataLocalProgramsPythonPython35libsite-packagestensorflowpythonframeworkops.py", line 2400, in _set_shapes_for_outputs
shapes = shape_func(op)
File "C:UsersASUSAppDataLocalProgramsPythonPython35libsite-packagestensorflowpythonframeworkops.py", line 2330, in call_with_requiring
return call_cpp_shape_fn(op, require_shape_fn=True)
File "C:UsersASUSAppDataLocalProgramsPythonPython35libsite-packagestensorflowpythonframeworkcommon_shapes.py", line 627, in call_cpp_shape_fn
require_shape_fn)
File "C:UsersASUSAppDataLocalProgramsPythonPython35libsite-packagestensorflowpythonframeworkcommon_shapes.py", line 691, in _call_cpp_shape_fn_impl
raise ValueError(err.message)
ValueError: Can not squeeze dim[1], expected a dimension of 1, got 3 for 'sparse_softmax_cross_entropy_loss/remove_squeezable_dimensions/Squeeze' (op: 'Squeeze') with input shapes: [100,3].
>>>
Here is my code:
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
#imports
import numpy as np
import tensorflow as tf
import glob
import cv2
import random
import matplotlib.pylab as plt
import pandas as pd
import sys as system
from mlxtend.preprocessing import one_hot
from sklearn import preprocessing
from sklearn.preprocessing import LabelEncoder
from sklearn.preprocessing import OneHotEncoder
tf.logging.set_verbosity(tf.logging.INFO)
def cnn_model_fn(features, labels, mode):
"""Model function for CNN"""
#Input Layer
input_layer = tf.reshape(features["x"], [-1,320,320,3])
#Convolutional Layer #1
conv1 = tf.layers.conv2d(
inputs = input_layer,
filters = 32,
kernel_size=[5,5],
padding = "same",
activation=tf.nn.relu)
#Pooling Layer #1
pool1 = tf.layers.max_pooling2d(inputs=conv1, pool_size=[2,2], strides=2)
#Convolutional Layer #2 and Pooling Layer #2
conv2 = tf.layers.conv2d(
inputs=pool1,
filters=64,
kernel_size=[5,5],
padding="same",
activation=tf.nn.relu)
pool2 = tf.layers.max_pooling2d(inputs=conv2, pool_size=[2,2], strides=2)
#Dense Layer
pool2_flat = tf.reshape(pool2, [-1,80*80*64])
dense = tf.layers.dense(inputs=pool2_flat, units=1024, activation=tf.nn.relu)
dropout = tf.layers.dropout(
inputs=dense, rate=0.4, training=mode == tf.estimator.ModeKeys.TRAIN)
#Logits Layer
logits = tf.layers.dense(inputs=dropout, units=3)
predictions = {
#Generate predictions (for PREDICT and EVAL mode)
"classes": tf.argmax(input=logits, axis=1),
#Add 'softmax_tensor' to the graph. It is used for PREDICT and by the
#'logging_hook'
"probabilities": tf.nn.softmax(logits, name="softmax_tensor")
}
if mode == tf.estimator.ModeKeys.PREDICT:
return tf.estimator.EstimatorSpec(mode=mode, predictions=predictions)
# Calculate Loss (for both TRAIN and EVAL modes
loss = tf.losses.sparse_softmax_cross_entropy(labels=labels, logits=logits)
# Configure the Training Op (for TRAIN mode)
if mode == tf.estimator.ModeKeys.TRAIN:
optimizer = tf.train.GradientDescentOptimizer(learning_rate=0.001)
train_op = optimizer.minimize(
loss=loss,
global_step=tf.train.get_global_step())
return tf.estimator.EstimatorSpec(mode=mode, loss=loss, train_op=train_op)
# Add evaluation metrics (for EVAL mode)
eval_metric_ops = {
"accuracy": tf.metrics.accuracy(
labels=labels, predictions=predictions["classes"])}
return tf.estimator.EstimatorSpec(
mode=mode, loss=loss,eval_metric_ops=eval_metric_ops)
def main(unused_argv):
'''
#Load training and eval data
mnist = tf.contrib.learn.datasets.load_dataset("mnist")
train_data = mnist.train.images
train_labels = np.asarray(mnist.train.labels, dtype=np.int32)
eval_data = mnist.test.images
eval_labels = np.asarray(mnist.test.labels, dtype=np.int32)
'''
#Load cats, dogs and cars image in local folder
X_data = []
files = glob.glob("data/cats/*.jpg")
for myFile in files:
image = cv2.imread (myFile)
imgR = cv2.resize(image, (320, 320))
imgNR = imgR/255
X_data.append(imgNR)
files = glob.glob("data/dogs/*.jpg")
for myFile in files:
image = cv2.imread (myFile)
imgR = cv2.resize(image, (320, 320))
imgNR = imgR/255
X_data.append(imgNR)
files = glob.glob ("data/cars/*.jpg")
for myFile in files:
image = cv2.imread (myFile)
imgR = cv2.resize(image, (320, 320))
imgNR = imgR/255
X_data.append (imgNR)
#print('X_data count:', len(X_data))
X_data_Val = []
files = glob.glob ("data/Validation/cats/*.jpg")
for myFile in files:
image = cv2.imread (myFile)
imgR = cv2.resize(image, (320, 320))
imgNR = imgR/255
X_data_Val.append (imgNR)
files = glob.glob ("data/Validation/dogs/*.jpg")
for myFile in files:
image = cv2.imread (myFile)
imgR = cv2.resize(image, (320, 320))
imgNR = imgR/255
X_data_Val.append (imgNR)
files = glob.glob ("data/Validation/cars/*.jpg")
for myFile in files:
image = cv2.imread (myFile)
imgR = cv2.resize(image, (320, 320))
imgNR = imgR/255
X_data_Val.append (imgNR)
#Feed One hot lables
Y_Label = np.zeros(shape=(300,1))
for el in range(0,100):
Y_Label[el]=[0]
for el in range(101,200):
Y_Label[el]=[1]
for el in range(201,300):
Y_Label[el]=[2]
onehot_encoder = OneHotEncoder(sparse=False)
#Y_Label_RS = Y_Label.reshape(len(Y_Label), 1)
Y_Label_Encode = onehot_encoder.fit_transform(Y_Label)
#print('Y_Label_Encode shape:', Y_Label_Encode.shape)
Y_Label_Val = np.zeros(shape=(30,1))
for el in range(0, 10):
Y_Label_Val[el]=[0]
for el in range(11, 20):
Y_Label_Val[el]=[1]
for el in range(21, 30):
Y_Label_Val[el]=[2]
#Y_Label_Val_RS = Y_Label_Val.reshape(len(Y_Label_Val), 1)
Y_Label_Val_Encode = onehot_encoder.fit_transform(Y_Label_Val)
#print('Y_Label_Val_Encode shape:', Y_Label_Val_Encode.shape)
train_data = np.array(X_data)
train_data = train_data.astype(np.float32)
train_labels = np.asarray(Y_Label_Encode, dtype=np.int32)
eval_data = np.array(X_data_Val)
eval_data = eval_data.astype(np.float32)
eval_labels = np.asarray(Y_Label_Val_Encode, dtype=np.int32)
print(train_data.shape)
print(train_labels.shape)
#Create the Estimator
mnist_classifier = tf.estimator.Estimator(
model_fn=cnn_model_fn, model_dir="/tmp/mnist_convnet_model")
# Set up logging for predictions
tensor_to_log = {"probabilities": "softmax_tensor"}
logging_hook = tf.train.LoggingTensorHook(
tensors=tensor_to_log, every_n_iter=50)
# Train the model
train_input_fn = tf.estimator.inputs.numpy_input_fn(
x={"x": train_data},
y=train_labels,
batch_size=100,
num_epochs=None,
shuffle=True)
mnist_classifier.train(
input_fn=train_input_fn,
#original steps are 20000
steps=1,
hooks=[logging_hook])
# Evaluate the model and print results
eval_input_fn = tf.estimator.inputs.numpy_input_fn(
x={"x": eval_data},
y=eval_labels,
num_epochs=1,
shuffle=False)
eval_results = mnist_classifier.evaluate(input_fn=eval_input_fn)
print(eval_results)
if __name__ == "__main__":
tf.app.run()
The error here is from tf.losses.sparse_softmax_cross_entropy(labels=labels, logits=logits).
The TensorFlow documentation clearly states that “labels vector must provide a single specific index for the true class for each row of logits”. So your labels vector must include only class-indices like 0,1,2 and not their respective one-hot-encodings like [1,0,0], [0,1,0], [0,0,1].
Reproducing the error to explain further:
import numpy as np
import tensorflow as tf
# Create random-array and assign as logits tensor
np.random.seed(12345)
logits = tf.convert_to_tensor(np.random.sample((4,4)))
print logits.get_shape() #[4,4]
# Create random-labels (Assuming only 4 classes)
labels = tf.convert_to_tensor(np.array([2, 2, 0, 1]))
loss_1 = tf.losses.sparse_softmax_cross_entropy(labels, logits)
sess = tf.Session()
sess.run(tf.global_variables_initializer())
print 'Loss: {}'.format(sess.run(loss_1)) # 1.44836854
# Now giving one-hot-encodings in place of class-indices for labels
wrong_labels = tf.convert_to_tensor(np.array([[0,0,1,0], [0,0,1,0], [1,0,0,0],[0,1,0,0]]))
loss_2 = tf.losses.sparse_softmax_cross_entropy(wrong_labels, logits)
# This should give you a similar error as soon as you define it
So try giving class-indices instead of one-hot encodings in your Y_Labels vector.
Hope this clears your doubt.
I have solved this error. The labels were in onehot encoding, so it was in dimension of [,10], rather than [,1]. So I used tf.argmax().
If you used Keras’ ImageDataGenerator, you can add class_mode="sparse" to obtain the correct levels:
train_datagen = keras.preprocessing.image.ImageDataGenerator(
rescale=1./255,
shear_range=0.2,
zoom_range=0.2,
horizontal_flip=True)
train_generator = train_datagen.flow_from_directory(
'data/train',
target_size=(150, 150),
batch_size=32,
class_mode="sparse")
Alternatively, you might be able to use softmax_cross_entropy, which seems to use onehot encoding for the labels.
i write the code that change [1,0,0], [0,1,0], [0,0,1] to 0,1,2.
import numpy as np
import tensorflow as tf
def change_to_right(wrong_labels):
right_labels=[]
for x in wrong_labels:
for i in range(0,len(wrong_labels[0])):
if x[i]==1:
right_labels.append(i)
return right_labels
wrong_labels =np.array([[0,0,1,0], [0,0,1,0], [1,0,0,0],[0,1,0,0]])
right_labels =tf.convert_to_tensor(np.array(change_to_right(wrong_labels)))
Changing
loss='sparse_categorical_crossentropy'
to
loss='categorical_crossentropy'
worked for me.
In simple english, your loss function should be categorical_crossentropy if you have applied labelbinarizer (for hot encoding) to your test data. If you have not hot encoded your test data, you should use’sparse_categorical_crossentropy’.
You can change to loss=’categorical_crossentropy’ for one hot encoding or the other option as mentioned earlier is
tf.losses.sparse_softmax_cross_entropy(labels, logits),