将 numba-scipy 用于特殊功能的问题

如何解决将 numba-scipy 用于特殊功能的问题

我想让 numba 与 scipy 特殊函数错误函数“scipy.special.erf”兼容，因为我在有限差分方案中使用这个函数，我相信使用 numba 可以加快速度。

为此，我使用 conda 安装了 numba-scipy 0.3.0，其中 should support the special function error function。

然后我构建了一个非常简单的示例，以查看是否可以使用调用 scipy.special.erf 的函数来运行 numba。代码如下：

from numba import jit
import scipy.special as sc

@jit
def f(x,y,N):
    for i in range(N):
        x +=sc.erf(y)
    return x


print(f(1,1,1))

我在这个问题的末尾复制了错误消息。一个问题显然与错误函数调用有关，但我对究竟出了什么问题以及如何修复它感到有些茫然。非常欢迎有关此问题的任何帮助。

错误信息如下：

test.py:4: NumbaWarning: 
Compilation is falling back to object mode WITH looplifting enabled because Function "f" Failed type inference due to: No implementation of function Function(<ufunc 'erf'>) found for signature:
 
 >>> erf(int64)
 
There are 2 candidate implementations:
  - Of which 2 did not match due to:
  Overload of function 'erf': File: numba_scipy/special/overloads.py: Line 9.
    With argument(s): '(int64)':
   No match.

During: resolving callee type: Function(<ufunc 'erf'>)
During: typing of call at test.py (7)


File "test.py",line 7:
def f(x,N):
    <source elided>
    for i in range(N):
        x +=sc.erf(y)
        ^

  @jit
test.py:4: NumbaWarning: 
Compilation is falling back to object mode WITHOUT looplifting enabled because Function "f" Failed type inference due to: cannot determine Numba type of <class 'numba.core.dispatcher.LiftedLoop'>

File "test.py",line 6:
def f(x,N):
    for i in range(N):
    ^

  @jit
/Users/*************/opt/anaconda3/lib/python3.8/site-packages/numba/core/object_mode_passes.py:177: NumbaWarning: Function "f" was compiled in object mode without forceobj=True,but has lifted loops.

File "test.py",N):
    for i in range(N):
    ^

  warnings.warn(errors.NumbaWarning(warn_msg,/Users/******************/opt/anaconda3/lib/python3.8/site-packages/numba/core/object_mode_passes.py:187: NumbaDeprecationWarning: 
Fall-back from the nopython compilation path to the object mode compilation path has been detected,this is deprecated behavIoUr.

For more information visit https://numba.pydata.org/numba-doc/latest/reference/deprecation.html#deprecation-of-object-mode-fall-back-behavIoUr-when-using-jit

File "test.py",N):
    for i in range(N):
    ^

  warnings.warn(errors.NumbaDeprecationWarning(msg,