def test_with_invalid_value_range(self):
values = [-1.0, 0.0, 1.5, 2.0, 5.0, 15]
with self.assertRaisesRegexp(
ValueError, "Shape must be rank 1 but is rank 0"):
histogram_ops.histogram_fixed_width(values, 1.0)
with self.assertRaisesRegexp(ValueError, "Dimension must be 2 but is 3"):
histogram_ops.histogram_fixed_width(values, [1.0, 2.0, 3.0])
def test_with_invalid_nbins(self):
values = [-1.0, 0.0, 1.5, 2.0, 5.0, 15]
with self.assertRaisesRegex(ValueError,
"Shape must be rank 0 but is rank 1"):
histogram_ops.histogram_fixed_width(values, [1.0, 5.0], nbins=[1, 2])
with self.assertRaisesRegex(ValueError, "Requires nbins > 0"):
histogram_ops.histogram_fixed_width(values, [1.0, 5.0], nbins=-5)
def test_with_invalid_nbins(self):
values = [-1.0, 0.0, 1.5, 2.0, 5.0, 15]
with self.assertRaisesRegexp(
ValueError, "Shape must be rank 0 but is rank 1"):
histogram_ops.histogram_fixed_width(values, [1.0, 5.0], nbins=[1, 2])
with self.assertRaisesRegexp(
ValueError, "Requires nbins > 0"):
histogram_ops.histogram_fixed_width(values, [1.0, 5.0], nbins=-5)
def test_two_updates_on_constant_input(self):
# Bins will be:
# (-inf, 1), [1, 2), [2, 3), [3, 4), [4, inf)
nbins = [5]
value_range = [0.0, 5.0]
new_values_1 = [-1.0, 0.0, 1.5, 2.0, 5.0, 15]
new_values_2 = [1.5, 4.5, 4.5, 4.5, 0.0, 0.0]
expected_bin_counts_1 = [2, 1, 1, 0, 2]
expected_bin_counts_2 = [4, 2, 1, 0, 5]
with self.test_session() as sess:
hist = variables.Variable(array_ops.zeros(nbins, dtype=dtypes.int32))
new_values = array_ops.placeholder(dtypes.float32, shape=[6])
hist_update = histogram_ops.histogram_fixed_width(hist, new_values,
value_range)
variables.initialize_all_variables().run()
updated_hist_array = sess.run(hist_update,
feed_dict={new_values: new_values_1})
# The new updated_hist_array is returned by the updating op.
# hist should contain the updated values.
self.assertAllClose(expected_bin_counts_1, updated_hist_array)
self.assertAllClose(expected_bin_counts_1, hist.eval())
updated_hist_array = sess.run(hist_update,
feed_dict={new_values: new_values_2})
self.assertAllClose(expected_bin_counts_2, updated_hist_array)
self.assertAllClose(expected_bin_counts_2, hist.eval())
def test_multiple_random_accumulating_updates_results_in_right_dist(self):
# Accumulate the updates in a new variable. Resultant
# histogram should be uniform. Use only 3 bins because with many bins it
# would be unlikely that all would be close to 1/n. If someone ever wants
# to test that, it would be better to check that the cdf was linear.
value_range = [1.0, 4.14159]
with self.test_session() as sess:
values = array_ops.placeholder(dtypes.float32, shape=[4, 4, 4])
hist = histogram_ops.histogram_fixed_width(
values, value_range, nbins=3, dtype=dtypes.int64)
hist_accum = variables.Variable(init_ops.zeros_initializer()(
[3], dtype=dtypes.int64))
hist_accum = hist_accum.assign_add(hist)
variables.global_variables_initializer().run()
for _ in range(100):
# Map the rv: U[0, 1] --> U[value_range[0], value_range[1]].
values_arr = (
value_range[0] +
(value_range[1] - value_range[0]) * self.rng.rand(4, 4, 4))
hist_accum_arr = sess.run(hist_accum, feed_dict={values: values_arr})
pmf = hist_accum_arr / float(hist_accum_arr.sum())
np.testing.assert_allclose(1 / 3, pmf, atol=0.02)
def test_multiple_random_3d_updates_results_in_right_dist(self):
# Update with uniform 3-D rvs. Resultant
# histogram should be uniform. Use only 3 bins because with many bins it
# would be unlikely that all would be close to 1/n. If someone ever wants
# to test that, it would be better to check that the cdf was linear.
nbins = [3]
value_range = [1.0, 4.14159]
with self.test_session() as sess:
hist = variables.Variable(array_ops.zeros(nbins, dtype=dtypes.int32))
new_values = array_ops.placeholder(dtypes.float32, shape=[4, 4, 4])
hist_update = histogram_ops.histogram_fixed_width(hist, new_values,
value_range)
variables.initialize_all_variables().run()
for _ in range(100):
# Map the rv: U[0, 1] --> U[value_range[0], value_range[1]].
new_values_arr = (
value_range[0] +
(value_range[1] - value_range[0]) * self.rng.rand(4, 4, 4))
# The new updated_hist_array is returned by the updating op.
# hist should contain the updated values.
updated_hist_array = sess.run(hist_update,
feed_dict={new_values: new_values_arr})
pmf = updated_hist_array / float(updated_hist_array.sum())
np.testing.assert_allclose(1 / 3, pmf, atol=0.02)
def test_two_updates_on_constant_input(self):
# Bins will be:
# (-inf, 1), [1, 2), [2, 3), [3, 4), [4, inf)
nbins = [5]
value_range = [0.0, 5.0]
new_values_1 = [-1.0, 0.0, 1.5, 2.0, 5.0, 15]
new_values_2 = [1.5, 4.5, 4.5, 4.5, 0.0, 0.0]
expected_bin_counts_1 = [2, 1, 1, 0, 2]
expected_bin_counts_2 = [4, 2, 1, 0, 5]
with self.test_session() as sess:
hist = variables.Variable(array_ops.zeros(nbins, dtype=dtypes.int32))
new_values = array_ops.placeholder(dtypes.float32, shape=[6])
hist_update = histogram_ops.histogram_fixed_width(hist, new_values,
value_range)
variables.initialize_all_variables().run()
updated_hist_array = sess.run(hist_update,
feed_dict={new_values: new_values_1})
# The new updated_hist_array is returned by the updating op.
# hist should contain the updated values.
self.assertAllClose(expected_bin_counts_1, updated_hist_array)
self.assertAllClose(expected_bin_counts_1, hist.eval())
updated_hist_array = sess.run(hist_update,
feed_dict={new_values: new_values_2})
self.assertAllClose(expected_bin_counts_2, updated_hist_array)
self.assertAllClose(expected_bin_counts_2, hist.eval())
def test_multiple_random_3d_updates_results_in_right_dist(self):
# Update with uniform 3-D rvs. Resultant
# histogram should be uniform. Use only 3 bins because with many bins it
# would be unlikely that all would be close to 1/n. If someone ever wants
# to test that, it would be better to check that the cdf was linear.
nbins = [3]
value_range = [1.0, 4.14159]
with self.test_session() as sess:
hist = variables.Variable(array_ops.zeros(nbins, dtype=dtypes.int32))
new_values = array_ops.placeholder(dtypes.float32, shape=[4, 4, 4])
hist_update = histogram_ops.histogram_fixed_width(hist, new_values,
value_range)
variables.initialize_all_variables().run()
for _ in range(100):
# Map the rv: U[0, 1] --> U[value_range[0], value_range[1]].
new_values_arr = (
value_range[0] +
(value_range[1] - value_range[0]) * self.rng.rand(4, 4, 4))
# The new updated_hist_array is returned by the updating op.
# hist should contain the updated values.
updated_hist_array = sess.run(hist_update,
feed_dict={new_values: new_values_arr})
pmf = updated_hist_array / float(updated_hist_array.sum())
np.testing.assert_allclose(1 / 3, pmf, atol=0.02)
def test_shape_inference(self):
value_range = [0.0, 5.0]
values = [[-1.0, 0.0, 1.5], [2.0, 5.0, 15]]
expected_bin_counts = [2, 1, 1, 0, 2]
placeholder = array_ops.placeholder(dtypes.int32)
with self.session(use_gpu=True):
hist = histogram_ops.histogram_fixed_width(values, value_range, nbins=5)
self.assertAllEqual(hist.shape.as_list(), (5,))
self.assertEqual(dtypes.int32, hist.dtype)
self.assertAllClose(expected_bin_counts, self.evaluate(hist))
hist = histogram_ops.histogram_fixed_width(
values, value_range, nbins=placeholder)
self.assertEquals(hist.shape.ndims, 1)
self.assertIs(hist.shape.dims[0].value, None)
self.assertEqual(dtypes.int32, hist.dtype)
self.assertAllClose(expected_bin_counts, hist.eval({placeholder: 5}))
def test_1d_float64_values(self):
# Bins will be:
# (-inf, 1), [1, 2), [2, 3), [3, 4), [4, inf)
value_range = np.float64([0.0, 5.0])
values = np.float64([-1.0, 0.0, 1.5, 2.0, 5.0, 15])
expected_bin_counts = [2, 1, 1, 0, 2]
with self.test_session(use_gpu=True):
hist = histogram_ops.histogram_fixed_width(values, value_range, nbins=5)
self.assertEqual(dtypes.int32, hist.dtype)
self.assertAllClose(expected_bin_counts, hist.eval())
def test_2d_values(self):
# Bins will be:
# (-inf, 1), [1, 2), [2, 3), [3, 4), [4, inf)
value_range = [0.0, 5.0]
values = [[-1.0, 0.0, 1.5], [2.0, 5.0, 15]]
expected_bin_counts = [2, 1, 1, 0, 2]
with self.session(use_gpu=True):
hist = histogram_ops.histogram_fixed_width(values, value_range, nbins=5)
self.assertEqual(dtypes.int32, hist.dtype)
self.assertAllClose(expected_bin_counts, self.evaluate(hist))
def test_empty_input_gives_all_zero_counts(self):
# Bins will be:
# (-inf, 1), [1, 2), [2, 3), [3, 4), [4, inf)
value_range = [0.0, 5.0]
values = []
expected_bin_counts = [0, 0, 0, 0, 0]
with self.session(use_gpu=True):
hist = histogram_ops.histogram_fixed_width(values, value_range, nbins=5)
self.assertEqual(dtypes.int32, hist.dtype)
self.assertAllClose(expected_bin_counts, self.evaluate(hist))
def test_1d_float64_values(self):
# Bins will be:
# (-inf, 1), [1, 2), [2, 3), [3, 4), [4, inf)
value_range = np.float64([0.0, 5.0])
values = np.float64([-1.0, 0.0, 1.5, 2.0, 5.0, 15])
expected_bin_counts = [2, 1, 1, 0, 2]
with self.test_session(use_gpu=True):
hist = histogram_ops.histogram_fixed_width(values, value_range, nbins=5)
self.assertEqual(dtypes.int32, hist.dtype)
self.assertAllClose(expected_bin_counts, hist.eval())
def test_1d_values_int64_output(self):
# Bins will be:
# (-inf, 1), [1, 2), [2, 3), [3, 4), [4, inf)
value_range = [0.0, 5.0]
values = [-1.0, 0.0, 1.5, 2.0, 5.0, 15]
expected_bin_counts = [2, 1, 1, 0, 2]
with self.session():
hist = histogram_ops.histogram_fixed_width(
values, value_range, nbins=5, dtype=dtypes.int64)
self.assertEqual(dtypes.int64, hist.dtype)
self.assertAllClose(expected_bin_counts, self.evaluate(hist))
def test_1d_values_int64_output(self):
# Bins will be:
# (-inf, 1), [1, 2), [2, 3), [3, 4), [4, inf)
value_range = [0.0, 5.0]
values = [-1.0, 0.0, 1.5, 2.0, 5.0, 15]
expected_bin_counts = [2, 1, 1, 0, 2]
with self.session(use_gpu=True):
hist = histogram_ops.histogram_fixed_width(
values, value_range, nbins=5, dtype=dtypes.int64)
self.assertEqual(dtypes.int64, hist.dtype)
self.assertAllClose(expected_bin_counts, hist.eval())
def _make_auc_histograms(boolean_labels, scores, score_range, nbins):
"""Create histogram tensors from one batch of labels/scores."""
with variable_scope.variable_scope(
None, 'make_auc_histograms', [boolean_labels, scores, nbins]):
# Histogram of scores for records in this batch with True label.
hist_true = histogram_ops.histogram_fixed_width(
array_ops.boolean_mask(scores, boolean_labels),
score_range,
nbins=nbins,
dtype=dtypes.int64,
name='hist_true')
# Histogram of scores for records in this batch with False label.
hist_false = histogram_ops.histogram_fixed_width(
array_ops.boolean_mask(scores, math_ops.logical_not(boolean_labels)),
score_range,
nbins=nbins,
dtype=dtypes.int64,
name='hist_false')
return hist_true, hist_false
def _make_auc_histograms(boolean_labels, scores, score_range, nbins):
"""Create histogram tensors from one batch of labels/scores."""
with variable_scope.variable_op_scope(
[boolean_labels, scores, nbins], None, 'make_auc_histograms'):
# Histogram of scores for records in this batch with True label.
hist_true = histogram_ops.histogram_fixed_width(
array_ops.boolean_mask(scores, boolean_labels),
score_range,
nbins=nbins,
dtype=dtypes.int64,
name='hist_true')
# Histogram of scores for records in this batch with False label.
hist_false = histogram_ops.histogram_fixed_width(
array_ops.boolean_mask(scores, math_ops.logical_not(boolean_labels)),
score_range,
nbins=nbins,
dtype=dtypes.int64,
name='hist_false')
return hist_true, hist_false
def test_one_update_on_constant_2d_input(self):
# Bins will be:
# (-inf, 1), [1, 2), [2, 3), [3, 4), [4, inf)
value_range = [0.0, 5.0]
values = [[-1.0, 0.0, 1.5], [2.0, 5.0, 15]]
expected_bin_counts = [2, 1, 1, 0, 2]
with self.test_session():
hist = histogram_ops.histogram_fixed_width(values, value_range, nbins=5)
# Hist should start "fresh" with every eval.
self.assertAllClose(expected_bin_counts, hist.eval())
self.assertAllClose(expected_bin_counts, hist.eval())
def test_empty_input_gives_all_zero_counts(self):
# Bins will be:
# (-inf, 1), [1, 2), [2, 3), [3, 4), [4, inf)
value_range = [0.0, 5.0]
values = []
expected_bin_counts = [0, 0, 0, 0, 0]
with self.test_session():
hist = histogram_ops.histogram_fixed_width(values, value_range, nbins=5)
# Hist should start "fresh" with every eval.
self.assertAllClose(expected_bin_counts, hist.eval())
self.assertAllClose(expected_bin_counts, hist.eval())
def histogram(self, x, value_range=None, nbins=None, name=None):
"""Return histogram of values.
Given the tensor `values`, this operation returns a rank 1 histogram
counting the number of entries in `values` that fell into every bin. The
bins are equal width and determined by the arguments `value_range` and
`nbins`.
Args:
x: 1D numeric `Tensor` of items to count.
value_range: Shape [2] `Tensor`. `new_values <= value_range[0]` will be
mapped to `hist[0]`, `values >= value_range[1]` will be mapped to
`hist[-1]`. Must be same dtype as `x`.
nbins: Scalar `int32 Tensor`. Number of histogram bins.
name: Python `str` name prefixed to Ops created by this class.
Returns:
counts: 1D `Tensor` of counts, i.e.,
`counts[i] = sum{ edges[i-1] <= values[j] < edges[i] : j }`.
edges: 1D `Tensor` characterizing intervals used for counting.
"""
with ops.name_scope(name, "histogram", [x]):
x = ops.convert_to_tensor(x, name="x")
if value_range is None:
value_range = [
math_ops.reduce_min(x), 1 + math_ops.reduce_max(x)
]
value_range = ops.convert_to_tensor(value_range,
name="value_range")
lo = value_range[0]
hi = value_range[1]
if nbins is None:
nbins = math_ops.to_int32(hi - lo)
delta = (hi - lo) / math_ops.cast(
nbins, dtype=value_range.dtype.base_dtype)
edges = math_ops.range(start=lo,
limit=hi,
delta=delta,
dtype=x.dtype.base_dtype)
counts = histogram_ops.histogram_fixed_width(
x, value_range=value_range, nbins=nbins)
return counts, edges
def test_one_update_on_constant_2d_input(self):
# Bins will be:
# (-inf, 1), [1, 2), [2, 3), [3, 4), [4, inf)
nbins = [5]
value_range = [0.0, 5.0]
new_values = [[-1.0, 0.0, 1.5], [2.0, 5.0, 15]]
expected_bin_counts = [2, 1, 1, 0, 2]
with self.test_session() as sess:
hist = variables.Variable(array_ops.zeros(nbins, dtype=dtypes.int32))
hist_update = histogram_ops.histogram_fixed_width(hist, new_values,
value_range)
variables.initialize_all_variables().run()
self.assertTrue(hist.dtype.is_compatible_with(hist_update.dtype))
updated_hist_array = sess.run(hist_update)
# The new updated_hist_array is returned by the updating op.
self.assertAllClose(expected_bin_counts, updated_hist_array)
# hist should contain updated values, but eval() should not change it.
self.assertAllClose(expected_bin_counts, hist.eval())
self.assertAllClose(expected_bin_counts, hist.eval())
def test_one_update_on_constant_2d_input(self):
# Bins will be:
# (-inf, 1), [1, 2), [2, 3), [3, 4), [4, inf)
nbins = [5]
value_range = [0.0, 5.0]
new_values = [[-1.0, 0.0, 1.5], [2.0, 5.0, 15]]
expected_bin_counts = [2, 1, 1, 0, 2]
with self.test_session() as sess:
hist = variables.Variable(array_ops.zeros(nbins, dtype=dtypes.int32))
hist_update = histogram_ops.histogram_fixed_width(hist, new_values,
value_range)
variables.initialize_all_variables().run()
self.assertTrue(hist.dtype.is_compatible_with(hist_update.dtype))
updated_hist_array = sess.run(hist_update)
# The new updated_hist_array is returned by the updating op.
self.assertAllClose(expected_bin_counts, updated_hist_array)
# hist should contain updated values, but eval() should not change it.
self.assertAllClose(expected_bin_counts, hist.eval())
self.assertAllClose(expected_bin_counts, hist.eval())
def test_two_updates_on_scalar_input(self):
# Bins will be:
# (-inf, 1), [1, 2), [2, 3), [3, 4), [4, inf)
value_range = [0.0, 5.0]
values_1 = 1.5
values_2 = 2.5
expected_bin_counts_1 = [0, 1, 0, 0, 0]
expected_bin_counts_2 = [0, 0, 1, 0, 0]
with self.test_session():
values = array_ops.placeholder(dtypes.float32, shape=[])
hist = histogram_ops.histogram_fixed_width(values, value_range, nbins=5)
# The values in hist should depend on the current feed and nothing else.
self.assertAllClose(
expected_bin_counts_2, hist.eval(feed_dict={values: values_2}))
self.assertAllClose(
expected_bin_counts_1, hist.eval(feed_dict={values: values_1}))
self.assertAllClose(
expected_bin_counts_1, hist.eval(feed_dict={values: values_1}))
self.assertAllClose(
expected_bin_counts_2, hist.eval(feed_dict={values: values_2}))
def histogram(self, x, value_range=None, nbins=None, name=None):
"""Return histogram of values.
Given the tensor `values`, this operation returns a rank 1 histogram
counting the number of entries in `values` that fell into every bin. The
bins are equal width and determined by the arguments `value_range` and
`nbins`.
Args:
x: 1D numeric `Tensor` of items to count.
value_range: Shape [2] `Tensor`. `new_values <= value_range[0]` will be
mapped to `hist[0]`, `values >= value_range[1]` will be mapped to
`hist[-1]`. Must be same dtype as `x`.
nbins: Scalar `int32 Tensor`. Number of histogram bins.
name: Python `str` name prefixed to Ops created by this class.
Returns:
counts: 1D `Tensor` of counts, i.e.,
`counts[i] = sum{ edges[i-1] <= values[j] < edges[i] : j }`.
edges: 1D `Tensor` characterizing intervals used for counting.
"""
with ops.name_scope(name, "histogram", [x]):
x = ops.convert_to_tensor(x, name="x")
if value_range is None:
value_range = [math_ops.reduce_min(x), 1 + math_ops.reduce_max(x)]
value_range = ops.convert_to_tensor(value_range, name="value_range")
lo = value_range[0]
hi = value_range[1]
if nbins is None:
nbins = math_ops.cast(hi - lo, dtypes.int32)
delta = (hi - lo) / math_ops.cast(
nbins, dtype=value_range.dtype.base_dtype)
edges = math_ops.range(
start=lo, limit=hi, delta=delta, dtype=x.dtype.base_dtype)
counts = histogram_ops.histogram_fixed_width(
x, value_range=value_range, nbins=nbins)
return counts, edges