from __future__ import annotations
import numpy as np
from ..utils import check
from .._tensor import _Tensor
from ..grad import operations, dag
[docs]
def min(input: _Tensor, other: _Tensor) -> _Tensor:
"""
Return the elementwise minimum of two tensors.
Parameters
----------
input : _Tensor
First input tensor.
other : _Tensor
Second input tensor.
Returns
-------
_Tensor
Elementwise minimum of ``input`` and ``other``.
See Also
--------
`torch.minimum <https://pytorch.org/docs/stable/generated/torch.minimum.html>`_
"""
if not isinstance(input, _Tensor) or not isinstance(other, _Tensor):
raise TypeError(f"Expected inputs to be eazygrad tensors, but got {type(input)} and {type(other)}.")
requires_grad = input.requires_grad or other.requires_grad
min_val = np.minimum(input._array, other._array)
result = _Tensor(min_val, requires_grad = requires_grad)
if requires_grad:
result.node_id = dag.create_node(
parents_id = [input.node_id, other.node_id],
operation = operations.Min(arr1=input._array, arr2=other._array),
result = result
)
return result
[docs]
def exp(input: _Tensor) -> _Tensor:
"""
Compute the elementwise exponential.
Parameters
----------
input : _Tensor
Input tensor.
Returns
-------
_Tensor
Tensor containing ``exp(input)``.
See Also
--------
`torch.exp <https://pytorch.org/docs/stable/generated/torch.exp.html>`_
"""
if not isinstance(input, _Tensor):
raise TypeError(f"Expected input to be an eazygrad tensor, got {type(input)}")
requires_grad = input.requires_grad
array = input._array
result = _Tensor(np.exp(array), requires_grad = requires_grad)
if requires_grad :
result.node_id = dag.create_node(parents_id = [input.node_id], operation = operations.Exp(arr=input._array), result = result)
return result
[docs]
def log(input: _Tensor) -> _Tensor:
"""
Compute the elementwise natural logarithm.
Parameters
----------
input : _Tensor
Input tensor. Values should be strictly positive.
Returns
-------
_Tensor
Tensor containing ``log(input)``.
See Also
--------
`torch.log <https://pytorch.org/docs/stable/generated/torch.log.html>`_
"""
if not isinstance(input, _Tensor):
raise TypeError(f"Expected input to be an eazygrad tensor, got {type(input)}")
requires_grad = input.requires_grad
array = input._array
result = _Tensor(np.log(array), requires_grad = requires_grad)
if requires_grad :
result.node_id = dag.create_node(parents_id = [input.node_id], operation = operations.Log(arr=input._array), result = result)
return result
[docs]
def cos(input: _Tensor) -> _Tensor:
"""
Compute the elementwise cosine.
Parameters
----------
input : _Tensor
Input tensor.
Returns
-------
_Tensor
Tensor containing ``cos(input)``.
See Also
--------
`torch.cos <https://pytorch.org/docs/stable/generated/torch.cos.html>`_
"""
if not isinstance(input, _Tensor):
raise TypeError(f"Expected input to be an eazygrad tensor, got {type(input)}")
requires_grad = input.requires_grad
result = _Tensor(np.cos(input._array), requires_grad = requires_grad)
if requires_grad :
result.node_id = dag.create_node(parents_id = [input.node_id], operation = operations.Cos(arr=input._array), result = result)
return result
[docs]
def sin(input: _Tensor) -> _Tensor:
"""
Compute the elementwise sine.
Parameters
----------
input : _Tensor
Input tensor.
Returns
-------
_Tensor
Tensor containing ``sin(input)``.
See Also
--------
`torch.sin <https://pytorch.org/docs/stable/generated/torch.sin.html>`_
"""
if not isinstance(input, _Tensor):
raise TypeError(f"Expected input to be an eazygrad tensor, got {type(input)}")
requires_grad = input.requires_grad
result = _Tensor(np.sin(input._array), requires_grad = requires_grad)
if requires_grad :
result.node_id = dag.create_node(parents_id = [input.node_id], operation = operations.Sin(arr=input._array), result = result)
return result
[docs]
def clip(input: _Tensor, low: float, high: float) -> _Tensor:
"""
Clip tensor values to a closed interval.
Parameters
----------
input : _Tensor
Input tensor.
low : scalar
Lower clipping bound.
high : scalar
Upper clipping bound.
Returns
-------
_Tensor
Clipped tensor.
See Also
--------
`torch.clamp <https://pytorch.org/docs/stable/generated/torch.clamp.html>`_
"""
if not isinstance(input, _Tensor):
raise TypeError(f"Expected input to be an eazygrad tensor, got {type(input)}")
if not check.is_scalar(low) or not check.is_scalar(high):
raise TypeError(f"Expected scalar bounds, got {type(low)} and {type(high)}")
if low > high:
raise ValueError(f"Expected low <= high, got low={low}, high={high}")
requires_grad = input.requires_grad
result = _Tensor(np.clip(input._array, low, high), requires_grad=requires_grad)
if requires_grad:
result.node_id = dag.create_node(
parents_id=[input.node_id],
operation=operations.Clip(arr=input._array, low=low, high=high),
result=result,
)
return result