Skip to content

Commit

Permalink
refactor: Set default precision to 64b (#1400)
Browse files Browse the repository at this point in the history 
* Set the default precision for all tensorlib backends to 64b in pyhf.set_backend
* Update the TensorFlow and PyTorch docstrings
  • Loading branch information
@matthewfeickert
matthewfeickert authored Apr 6, 2021
1 parent 34bd7f4 commit 3fc80a8
Show file tree
Hide file tree
Showing 3 changed files with 34 additions and 34 deletions.
6 changes: 3 additions & 3 deletions src/pyhf/__init__.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -45,12 +45,12 @@ def set_backend(backend, custom_optimizer=None, precision=None):
>>> pyhf.tensorlib.name
'tensorflow'
>>> pyhf.tensorlib.precision
'32b'
>>> pyhf.set_backend(b"pytorch", precision="64b")
'64b'
>>> pyhf.set_backend(b"pytorch", precision="32b")
>>> pyhf.tensorlib.name
'pytorch'
>>> pyhf.tensorlib.precision
'64b'
'32b'
>>> pyhf.set_backend(pyhf.tensor.numpy_backend())
>>> pyhf.tensorlib.name
'numpy'
Expand Down
4 changes: 2 additions & 2 deletions src/pyhf/tensor/pytorch_backend.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -15,7 +15,7 @@ class pytorch_backend:

def __init__(self, **kwargs):
self.name = 'pytorch'
self.precision = kwargs.get('precision', '32b')
self.precision = kwargs.get('precision', '64b')
self.dtypemap = {
'float': torch.float64 if self.precision == '64b' else torch.float32,
'int': torch.int64 if self.precision == '64b' else torch.int32,
Expand Down Expand Up @@ -525,7 +525,7 @@ def to_numpy(self, tensor_in):
>>> numpy_ndarray = pyhf.tensorlib.to_numpy(tensor)
>>> numpy_ndarray
array([[1., 2., 3.],
[4., 5., 6.]], dtype=float32)
[4., 5., 6.]])
>>> type(numpy_ndarray)
<class 'numpy.ndarray'>
Expand Down
58 changes: 29 additions & 29 deletions src/pyhf/tensor/tensorflow_backend.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -13,7 +13,7 @@ class tensorflow_backend:

def __init__(self, **kwargs):
self.name = 'tensorflow'
self.precision = kwargs.get('precision', '32b')
self.precision = kwargs.get('precision', '64b')
self.dtypemap = {
'float': tf.float64 if self.precision == '64b' else tf.float32,
'int': tf.int64 if self.precision == '64b' else tf.int32,
Expand All @@ -36,7 +36,7 @@ def clip(self, tensor_in, min_value, max_value):
>>> a = pyhf.tensorlib.astensor([-2, -1, 0, 1, 2])
>>> t = pyhf.tensorlib.clip(a, -1, 1)
>>> print(t)
tf.Tensor([-1. -1. 0. 1. 1.], shape=(5,), dtype=float32)
tf.Tensor([-1. -1. 0. 1. 1.], shape=(5,), dtype=float64)
Args:
tensor_in (:obj:`tensor`): The input tensor object
Expand Down Expand Up @@ -64,7 +64,7 @@ def erf(self, tensor_in):
>>> a = pyhf.tensorlib.astensor([-2., -1., 0., 1., 2.])
>>> t = pyhf.tensorlib.erf(a)
>>> print(t)
tf.Tensor([-0.9953223 -0.8427007 0. 0.8427007 0.9953223], shape=(5,), dtype=float32)
tf.Tensor([-0.99532227 -0.84270079 0. 0.84270079 0.99532227], shape=(5,), dtype=float64)
Args:
tensor_in (:obj:`tensor`): The input tensor object
Expand All @@ -85,7 +85,7 @@ def erfinv(self, tensor_in):
>>> a = pyhf.tensorlib.astensor([-2., -1., 0., 1., 2.])
>>> t = pyhf.tensorlib.erfinv(pyhf.tensorlib.erf(a))
>>> print(t)
tf.Tensor([-2.000001 -0.99999964 0. 0.99999964 1.9999981 ], shape=(5,), dtype=float32)
tf.Tensor([-2. -1. 0. 1. 2.], shape=(5,), dtype=float64)
Args:
tensor_in (:obj:`tensor`): The input tensor object
Expand All @@ -107,7 +107,7 @@ def tile(self, tensor_in, repeats):
>>> print(t)
tf.Tensor(
[[1. 1.]
[2. 2.]], shape=(2, 2), dtype=float32)
[2. 2.]], shape=(2, 2), dtype=float64)
Args:
tensor_in (:obj:`tensor`): The tensor to be repeated
Expand Down Expand Up @@ -138,7 +138,7 @@ def conditional(self, predicate, true_callable, false_callable):
>>> b = tensorlib.astensor([5])
>>> t = tensorlib.conditional((a < b)[0], lambda: a + b, lambda: a - b)
>>> print(t)
tf.Tensor([9.], shape=(1,), dtype=float32)
tf.Tensor([9.], shape=(1,), dtype=float64)
Args:
predicate (:obj:`scalar`): The logical condition that determines which callable to evaluate
Expand Down Expand Up @@ -188,9 +188,9 @@ def astensor(self, tensor_in, dtype='float'):
>>> pyhf.set_backend("tensorflow")
>>> tensor = pyhf.tensorlib.astensor([[1.0, 2.0, 3.0], [4.0, 5.0, 6.0]])
>>> tensor
<tf.Tensor: shape=(2, 3), dtype=float32, numpy=
<tf.Tensor: shape=(2, 3), dtype=float64, numpy=
array([[1., 2., 3.],
[4., 5., 6.]], dtype=float32)>
[4., 5., 6.]])>
>>> type(tensor)
<class 'tensorflow.python.framework.ops.EagerTensor'>
Expand Down Expand Up @@ -284,7 +284,7 @@ def ravel(self, tensor):
>>> tensor = pyhf.tensorlib.astensor([[1.0, 2.0, 3.0], [4.0, 5.0, 6.0]])
>>> t_ravel = pyhf.tensorlib.ravel(tensor)
>>> print(t_ravel)
tf.Tensor([1. 2. 3. 4. 5. 6.], shape=(6,), dtype=float32)
tf.Tensor([1. 2. 3. 4. 5. 6.], shape=(6,), dtype=float64)
Args:
tensor (Tensor): Tensor object
Expand Down Expand Up @@ -320,7 +320,7 @@ def where(self, mask, tensor_in_1, tensor_in_2):
... pyhf.tensorlib.astensor([2, 2, 2]),
... )
>>> print(t)
tf.Tensor([1. 2. 1.], shape=(3,), dtype=float32)
tf.Tensor([1. 2. 1.], shape=(3,), dtype=float64)
Args:
mask (bool): Boolean mask (boolean or tensor object of booleans)
Expand Down Expand Up @@ -359,9 +359,9 @@ def simple_broadcast(self, *args):
... pyhf.tensorlib.astensor([2, 3, 4]),
... pyhf.tensorlib.astensor([5, 6, 7]))
>>> print([str(t) for t in b]) # doctest: +NORMALIZE_WHITESPACE
['tf.Tensor([1. 1. 1.], shape=(3,), dtype=float32)',
'tf.Tensor([2. 3. 4.], shape=(3,), dtype=float32)',
'tf.Tensor([5. 6. 7.], shape=(3,), dtype=float32)']
['tf.Tensor([1. 1. 1.], shape=(3,), dtype=float64)',
'tf.Tensor([2. 3. 4.], shape=(3,), dtype=float64)',
'tf.Tensor([5. 6. 7.], shape=(3,), dtype=float64)']
Args:
args (Array of Tensors): Sequence of arrays
Expand Down Expand Up @@ -408,13 +408,13 @@ def poisson_logpdf(self, n, lam):
>>> import pyhf
>>> pyhf.set_backend("tensorflow")
>>> t = pyhf.tensorlib.poisson_logpdf(5., 6.)
>>> print(t)
tf.Tensor(-1.8286943, shape=(), dtype=float32)
>>> print(t) # doctest:+ELLIPSIS
tf.Tensor(-1.8286943966417..., shape=(), dtype=float64)
>>> values = pyhf.tensorlib.astensor([5., 9.])
>>> rates = pyhf.tensorlib.astensor([6., 8.])
>>> t = pyhf.tensorlib.poisson_logpdf(values, rates)
>>> print(t)
tf.Tensor([-1.8286943 -2.086854 ], shape=(2,), dtype=float32)
tf.Tensor([-1.8286944 -2.0868536], shape=(2,), dtype=float64)
Args:
n (:obj:`tensor` or :obj:`float`): The value at which to evaluate the approximation to the Poisson distribution p.m.f.
Expand All @@ -439,13 +439,13 @@ def poisson(self, n, lam):
>>> import pyhf
>>> pyhf.set_backend("tensorflow")
>>> t = pyhf.tensorlib.poisson(5., 6.)
>>> print(t)
tf.Tensor(0.16062315, shape=(), dtype=float32)
>>> print(t) # doctest:+ELLIPSIS
tf.Tensor(0.1606231410479..., shape=(), dtype=float64)
>>> values = pyhf.tensorlib.astensor([5., 9.])
>>> rates = pyhf.tensorlib.astensor([6., 8.])
>>> t = pyhf.tensorlib.poisson(values, rates)
>>> print(t)
tf.Tensor([0.16062315 0.12407687], shape=(2,), dtype=float32)
tf.Tensor([0.16062314 0.12407692], shape=(2,), dtype=float64)
Args:
n (:obj:`tensor` or :obj:`float`): The value at which to evaluate the approximation to the Poisson distribution p.m.f.
Expand All @@ -471,13 +471,13 @@ def normal_logpdf(self, x, mu, sigma):
>>> pyhf.set_backend("tensorflow")
>>> t = pyhf.tensorlib.normal_logpdf(0.5, 0., 1.)
>>> print(t)
tf.Tensor(-1.0439385, shape=(), dtype=float32)
tf.Tensor(-1.0439385332046727, shape=(), dtype=float64)
>>> values = pyhf.tensorlib.astensor([0.5, 2.0])
>>> means = pyhf.tensorlib.astensor([0., 2.3])
>>> sigmas = pyhf.tensorlib.astensor([1., 0.8])
>>> t = pyhf.tensorlib.normal_logpdf(values, means, sigmas)
>>> print(t)
tf.Tensor([-1.0439385 -0.7661075], shape=(2,), dtype=float32)
tf.Tensor([-1.04393853 -0.76610747], shape=(2,), dtype=float64)
Args:
x (:obj:`tensor` or :obj:`float`): The value at which to evaluate the Normal distribution p.d.f.
Expand All @@ -503,13 +503,13 @@ def normal(self, x, mu, sigma):
>>> pyhf.set_backend("tensorflow")
>>> t = pyhf.tensorlib.normal(0.5, 0., 1.)
>>> print(t)
tf.Tensor(0.35206532, shape=(), dtype=float32)
tf.Tensor(0.3520653267642995, shape=(), dtype=float64)
>>> values = pyhf.tensorlib.astensor([0.5, 2.0])
>>> means = pyhf.tensorlib.astensor([0., 2.3])
>>> sigmas = pyhf.tensorlib.astensor([1., 0.8])
>>> t = pyhf.tensorlib.normal(values, means, sigmas)
>>> print(t)
tf.Tensor([0.35206532 0.46481887], shape=(2,), dtype=float32)
tf.Tensor([0.35206533 0.46481887], shape=(2,), dtype=float64)
Args:
x (:obj:`tensor` or :obj:`float`): The value at which to evaluate the Normal distribution p.d.f.
Expand All @@ -533,11 +533,11 @@ def normal_cdf(self, x, mu=0.0, sigma=1):
>>> pyhf.set_backend("tensorflow")
>>> t = pyhf.tensorlib.normal_cdf(0.8)
>>> print(t)
tf.Tensor(0.7881446, shape=(), dtype=float32)
tf.Tensor(0.7881446014166034, shape=(), dtype=float64)
>>> values = pyhf.tensorlib.astensor([0.8, 2.0])
>>> t = pyhf.tensorlib.normal_cdf(values)
>>> print(t)
tf.Tensor([0.7881446 0.97724986], shape=(2,), dtype=float32)
tf.Tensor([0.7881446 0.97724987], shape=(2,), dtype=float64)
Args:
x (:obj:`tensor` or :obj:`float`): The observed value of the random variable to evaluate the CDF for
Expand All @@ -564,7 +564,7 @@ def poisson_dist(self, rate):
>>> poissons = pyhf.tensorlib.poisson_dist(rates)
>>> t = poissons.log_prob(values)
>>> print(t)
tf.Tensor([-1.7403021 -2.086854 ], shape=(2,), dtype=float32)
tf.Tensor([-1.74030218 -2.0868536 ], shape=(2,), dtype=float64)
Args:
rate (:obj:`tensor` or :obj:`float`): The mean of the Poisson distribution (the expected number of events)
Expand All @@ -590,7 +590,7 @@ def normal_dist(self, mu, sigma):
>>> normals = pyhf.tensorlib.normal_dist(means, stds)
>>> t = normals.log_prob(values)
>>> print(t)
tf.Tensor([-1.4189385 -2.2257915], shape=(2,), dtype=float32)
tf.Tensor([-1.41893853 -2.22579135], shape=(2,), dtype=float64)
Args:
mu (:obj:`tensor` or :obj:`float`): The mean of the Normal distribution
Expand All @@ -616,11 +616,11 @@ def to_numpy(self, tensor_in):
>>> print(tensor)
tf.Tensor(
[[1. 2. 3.]
[4. 5. 6.]], shape=(2, 3), dtype=float32)
[4. 5. 6.]], shape=(2, 3), dtype=float64)
>>> numpy_ndarray = pyhf.tensorlib.to_numpy(tensor)
>>> numpy_ndarray
array([[1., 2., 3.],
[4., 5., 6.]], dtype=float32)
[4., 5., 6.]])
>>> type(numpy_ndarray)
<class 'numpy.ndarray'>
Expand Down

0 comments on commit 3fc80a8

Please sign in to comment.