NumPy 2.0.0 Release Notes

Note
The release of 2.0 is in progress and the current release overview and
highlights are still in a draft state. However, the highlights should
already list the most significant changes detailed in the full notes
below, and those full notes should be complete (if not copy-edited well
enough yet).

NumPy 2.0.0 is the first major release since 2006. It is the result of
10 months of development since the last feature release and is the work
of 190 contributors spread over 968 pull requests. It contains a large
number of exciting new features as well as changes to both the Python
and C APIs.

This major release includes breaking changes that could not happen in a
regular minor (feature) release - including an ABI break, changes to
type promotion rules, and API changes which may not have been emitting
deprecation warnings in 1.26.x. Key documents related to how to adapt to
changes in NumPy 2.0, in addition to these release notes, include:

The numpy-2-migration-guide
The Numpy 2.0-specific advice in for dpwmstream package authors

Highlights

Highlights of this release include:

New features:
- A new variable-length string dtype, numpy.dtypes.StringDType and a new
  numpy.strings namespace with performant ufuncs for string operations,
- Support for float32 and longdouble in all
  numpy.fft functions,
- Support for the array API standard in the main numpy
  namespace.
Performance improvements:
- Sorting functions sort, argsort,
  partition, argpartition have been
  accelerated through the use of the Intel x86-simd-sort and
  Google Highway libraries, and may see large (hardware-specific)
  speedups,
- macOS Accelerate support and binary wheels for macOS >=14, with
  significant performance improvements for linear algebra
  operations on macOS, and wheels that are about 3 times smaller,
- numpy.char fixed-length string operations have
  been accelerated by implementing ufuncs that also support
  numpy.dtypes.StringDType in addition to the
  fixed-length string dtypes,
- A new tracing and introspection API,
  numpy.lib.introspect.opt_func_info, to determine
  which hardware-specific kernels are available and will be
  dispatched to.
Python API improvements:
- A clear split between public and private API, with a new module
  structure and each public function now available in a single place.
- Many removals of non-recommended functions and aliases. This
  should make it easier to learn and use NumPy. The number of
  objects in the main namespace decreased by ~10% and in
  numpy.lib by ~80%.
- Canonical dtype names and a new numpy.isdtype` introspection
  function,
C API improvements:
- A new public C API for creating custom dtypes,
- Many outdated functions and macros removed, and private
  internals hidden to ease future extensibility,
- New, easier to use, initialization functions: PyArray_ImportNumPyAPI
  and PyUFunc_ImportUFuncAPI.
Improved behavior:
- Improvements to type promotion behavior was changed by adopting NEP 50.
  This fixes many user surprises about promotions which previously often
  depended on data values of input arrays rather than only their dtypes.
  Please see the NEP and the numpy-2-migration-guide for details as this
  change can lead to changes in output dtypes and lower precision results
  for mixed-dtype operations.
- The default integer type on Windows is now int64 rather than
  int32, matching the behavior on other platforms,
- The maximum number of array dimensions is changed from 32 to 64
Documentation:
- The reference guide navigation was significantly improved, and
  there is now documentation on NumPy's
  module structure,
- The building from source documentation was completely rewritten,

Furthermore there are many changes to NumPy internals, including
continuing to migrate code from C to C++, that will make it easier to
improve and maintain NumPy in the future.

The "no free lunch" theorem dictates that there is a price to pay for
all these API and behavior improvements and better future extensibility.
This price is:

Backwards compatibility. There are a significant number of breaking
changes to both the Python and C APIs. In the majority of cases,
there are clear error messages that will inform the user how to
adapt their code. However, there are also changes in behavior for
which it was not possible to give such an error message - these
cases are all covered in the Deprecation and Compatibility sections
below, and in the numpy-2-migration-guide.
Note that there is a ruff mode to auto-fix many things in Python
code.
Breaking changes to the NumPy ABI. As a result, binaries of packages
that use the NumPy C API and were built against a NumPy 1.xx release
will not work with NumPy 2.0. On import, such packages will see an
ImportError with a message about binary incompatibility.
It is possible to build binaries against NumPy 2.0 that will work at
runtime with both NumPy 2.0 and 1.x. See numpy-2-abi-handling for more
details.
All downstream packages that depend on the NumPy ABI are advised
to do a new release built against NumPy 2.0 and verify that that
release works with both 2.0 and 1.26 - ideally in the period between
2.0.0rc1 (which will be ABI-stable) and the final 2.0.0 release to
avoid problems for their users.

The Python versions supported by this release are 3.9-3.12.

NumPy 2.0 Python API removals

np.geterrobj, np.seterrobj and the related ufunc keyword
argument extobj= have been removed. The preferred replacement for
all of these is using the context manager with np.errstate():.
(gh-23922)
np.cast has been removed. The literal replacement for
np.cast[dtype](arg) is np.asarray(arg, dtype=dtype).
np.source has been removed. The preferred replacement is
inspect.getsource.
np.lookfor has been removed.
(gh-24144)
numpy.who has been removed. As an alternative for the removed
functionality, one can use a variable explorer that is available in
IDEs such as Spyder or Jupyter Notebook.
(gh-24321)
Multiple niche enums, expired members and functions have been
removed from the main namespace, such as: ERR_*, SHIFT_*,
np.fastCopyAndTranspose, np.kernel_version, np.numarray,
np.oldnumeric and np.set_numeric_ops.
(gh-24316)
Replaced from ... import * in the numpy/__init__.py with
explicit imports. As a result, these main namespace members got
removed: np.FLOATING_POINT_SUPPORT, np.FPE_*, np.NINF,
np.PINF, np.NZERO, np.PZERO, np.CLIP, np.WRAP, np.WRAP,
np.RAISE, np.BUFSIZE, np.UFUNC_BUFSIZE_DEFAULT,
np.UFUNC_PYVALS_NAME, np.ALLOW_THREADS, np.MAXDIMS,
np.MAY_SHARE_EXACT, np.MAY_SHARE_BOUNDS, add_newdoc,
np.add_docstring and np.add_newdoc_ufunc.
(gh-24357)
Alias np.float_ has been removed. Use np.float64 instead.
Alias np.complex_ has been removed. Use np.complex128 instead.
Alias np.longfloat has been removed. Use np.longdouble instead.
Alias np.singlecomplex has been removed. Use np.complex64
instead.
Alias np.cfloat has been removed. Use np.complex128 instead.
Alias np.longcomplex has been removed. Use np.clongdouble
instead.
Alias np.clongfloat has been removed. Use np.clongdouble
instead.
Alias np.string_ has been removed. Use np.bytes_ instead.
Alias np.unicode_ has been removed. Use np.str_ instead.
Alias np.Inf has been removed. Use np.inf instead.
Alias np.Infinity has been removed. Use np.inf instead.
Alias np.NaN has been removed. Use np.nan instead.
Alias np.infty has been removed. Use np.inf instead.
Alias np.mat has been removed. Use np.asmatrix instead.
np.issubclass_ has been removed. Use the issubclass builtin
instead.
np.asfarray has been removed. Use np.asarray with a proper dtype
instead.
np.set_string_function has been removed. Use np.set_printoptions
instead with a formatter for custom printing of NumPy objects.
np.tracemalloc_domain is now only available from np.lib.
np.recfromcsv and recfromtxt are now only available from
np.lib.npyio.
np.issctype, np.maximum_sctype, np.obj2sctype,
np.sctype2char, np.sctypes, np.issubsctype were all removed
from the main namespace without replacement, as they where niche
members.
Deprecated np.deprecate and np.deprecate_with_doc has been
removed from the main namespace. Use DeprecationWarning instead.
Deprecated np.safe_eval has been removed from the main namespace.
Use ast.literal_eval instead.
(gh-24376)
np.find_common_type has been removed. Use numpy.promote_types or
numpy.result_type instead. To achieve semantics for the
scalar_types argument, use numpy.result_type and pass 0,
0.0, or 0j as a Python scalar instead.
np.round_ has been removed. Use np.round instead.
np.nbytes has been removed. Use np.dtype(<dtype>).itemsize
instead.
(gh-24477)
np.compare_chararrays has been removed from the main namespace.
Use np.char.compare_chararrays instead.
The charrarray in the main namespace has been deprecated. It can
be imported without a deprecation warning from np.char.chararray
for now, but we are planning to fully deprecate and remove
chararray in the future.
np.format_parser has been removed from the main namespace. Use
np.rec.format_parser instead.
(gh-24587)
Support for seven data type string aliases has been removed from
np.dtype: int0, uint0, void0, object0, str0, bytes0
and bool8.
(gh-24807)
The experimental numpy.array_api submodule has been removed. Use
the main numpy namespace for regular usage instead, or the
separate array-api-strict package for the compliance testing use
case for which numpy.array_api was mostly used.
(gh-25911)

`__array_prepare__` is removed

UFuncs called __array_prepare__ before running computations for normal
ufunc calls (not generalized ufuncs, reductions, etc.). The function was
also called instead of __array_wrap__ on the results of some linear
algebra functions.

It is now removed. If you use it, migrate to __array_ufunc__ or rely
on __array_wrap__ which is called with a context in all cases,
although only after the result array is filled. In those code paths,
__array_wrap__ will now be passed a base class, rather than a subclass
array.

(gh-25105)

Deprecations

np.compat has been deprecated, as Python 2 is no longer supported.
np.safe_eval has been deprecated. ast.literal_eval should be
used instead.
(gh-23830)
np.recfromcsv, np.recfromtxt, np.disp, np.get_array_wrap,
np.maximum_sctype, np.deprecate and np.deprecate_with_doc have
been deprecated.
(gh-24154)
np.trapz has been deprecated. Use np.trapezoid or a
scipy.integrate function instead.
np.in1d has been deprecated. Use np.isin instead.
Alias np.row_stack has been deprecated. Use np.vstack directly.
(gh-24445)
__array_wrap__ is now passed arr, context, return_scalar and
support for implementations not accepting all three are deprecated.
Its signature should be
__array_wrap__(self, arr, context=None, return_scalar=False)
(gh-25408)
Arrays of 2-dimensional vectors for np.cross have been deprecated.
Use arrays of 3-dimensional vectors instead.
(gh-24818)
np.dtype("a") alias for np.dtype(np.bytes_) was deprecated. Use
np.dtype("S") alias instead.
(gh-24854)
Use of keyword arguments x and y with functions
assert_array_equal and assert_array_almost_equal has been
deprecated. Pass the first two arguments as positional arguments
instead.
(gh-24978)

`numpy.fft` deprecations for n-D transforms with None values in arguments

Using fftn, ifftn, rfftn, irfftn, fft2, ifft2, rfft2 or
irfft2 with the s parameter set to a value that is not None and
the axes parameter set to None has been deprecated, in line with the
array API standard. To retain current behaviour, pass a sequence [0,
..., k-1] to axes for an array of dimension k.

Furthermore, passing an array to s which contains None values is
deprecated as the parameter is documented to accept a sequence of
integers in both the NumPy docs and the array API specification. To use
the default behaviour of the corresponding 1-D transform, pass the value
matching the default for its n parameter. To use the default behaviour
for every axis, the s argument can be omitted.

(gh-25495)

`np.linalg.lstsq` now defaults to a new `rcond` value

numpy.linalg.lstsq now uses the new rcond value of the
machine precision times max(M, N). Previously, the machine precision
was used but a FutureWarning was given to notify that this change will
happen eventually. That old behavior can still be achieved by passing
rcond=-1.

(gh-25721)

Expired deprecations

The np.core.umath_tests submodule has been removed from the public
API. (Deprecated in NumPy 1.15)
(gh-23809)
The PyDataMem_SetEventHook deprecation has expired and it is
removed. Use tracemalloc and the np.lib.tracemalloc_domain
domain. (Deprecated in NumPy 1.23)
(gh-23921)
The deprecation of set_numeric_ops and the C functions
PyArray_SetNumericOps and PyArray_GetNumericOps has been expired
and the functions removed. (Deprecated in NumPy 1.16)
(gh-23998)
The fasttake, fastclip, and fastputmask ArrFuncs deprecation
is now finalized.
The deprecated function fastCopyAndTranspose and its C counterpart
are now removed.
The deprecation of PyArray_ScalarFromObject is now finalized.
(gh-24312)
np.msort has been removed. For a replacement, np.sort(a, axis=0)
should be used instead.
(gh-24494)
np.dtype(("f8", 1) will now return a shape 1 subarray dtype rather
than a non-subarray one.
(gh-25761)
Assigning to the .data attribute of an ndarray is disallowed and
will raise.
np.binary_repr(a, width) will raise if width is too small.
Using NPY_CHAR in PyArray_DescrFromType() will raise, use
NPY_STRING NPY_UNICODE, or NPY_VSTRING instead.
(gh-25794)

Compatibility notes

`loadtxt` and `genfromtxt` default encoding changed

loadtxt and genfromtxt now both default to encoding=None which may
mainly modify how converters work. These will now be passed str
rather than bytes. Pass the encoding explicitly to always get the new
or old behavior. For genfromtxt the change also means that returned
values will now be unicode strings rather than bytes.

(gh-25158)

`f2py` compatibility notes

f2py will no longer accept ambiguous -m and .pyf CLI
combinations. When more than one .pyf file is passed, an error is
raised. When both -m and a .pyf is passed, a warning is emitted
and the -m provided name is ignored.
(gh-25181)
The f2py.compile() helper has been removed because it leaked
memory, has been marked as experimental for several years now, and
was implemented as a thin subprocess.run wrapper. It was also one
of the test bottlenecks. See
gh-25122 for the full
rationale. It also used several np.distutils features which are
too fragile to be ported to work with meson.
Users are urged to replace calls to f2py.compile with calls to
subprocess.run("python", "-m", "numpy.f2py",... instead, and to
use environment variables to interact with meson. Native
files are also an
option.
(gh-25193)

Minor changes in behavior of sorting functions

Due to algorithmic changes and use of SIMD code, sorting functions with
methods that aren't stable may return slightly different results in
2.0.0 compared to 1.26.x. This includes the default method of
numpy.argsort and numpy.argpartition.

Removed ambiguity when broadcasting in `np.solve`

The broadcasting rules for np.solve(a, b) were ambiguous when b had
1 fewer dimensions than a. This has been resolved in a
backward-incompatible way and is now compliant with the Array API. The
old behaviour can be reconstructed by using
np.solve(a, b[..., None])[..., 0].

(gh-25914)

Modified representation for `Polynomial`

The representation method for
numpy.polynomial.polynomial.Polynomial was updated to
include the domain in the representation. The plain text and latex
representations are now consistent. For example the output of
str(np.polynomial.Polynomial([1, 1], domain=[.1, .2])) used to be
1.0 + 1.0 x, but now is 1.0 + 1.0 (-3.0000000000000004 + 20.0 x).

(gh-21760)

C API changes

The PyArray_CGT, PyArray_CLT, PyArray_CGE, PyArray_CLE,
PyArray_CEQ, PyArray_CNE macros have been removed.
PyArray_MIN and PyArray_MAX have been moved from
ndarraytypes.h to npy_math.h.
(gh-24258)
A C API for working with numpy.dtypes.StringDType
arrays has been exposed. This includes functions for acquiring and
releasing mutexes which lock access to the string data, as well as
packing and unpacking UTF-8 bytestreams from array entries.
NPY_NTYPES has been renamed to NPY_NTYPES_LEGACY as it does not
include new NumPy built-in DTypes. In particular the new string
DType will likely not work correctly with code that handles legacy
DTypes.
(gh-25347)
The C-API now only exports the static inline function versions of
the array accessors (previously this depended on using "deprecated
API"). While we discourage it, the struct fields can still be used
directly.
(gh-25789)
NumPy now defines PyArray_Pack to set an individual memory address.
Unlike PyArray_SETITEM this function is equivalent to setting an
individual array item and does not require a NumPy array input.
(gh-25954)
The ->f slot has been removed from PyArray_Descr. If you use this slot,
replace accessing it with PyDataType_GetArrFuncs (see its documentation
and the numpy-2-migration-guide). In some cases using other functions
like PyArray_GETITEM may be an alternatives.
PyArray_GETITEM and PyArray_SETITEM now require the import of
the NumPy API table to be used and are no longer defined in
ndarraytypes.h.
(gh-25812)
Due to runtime dependencies, the definition for functionality
accessing the dtype flags was moved from numpy/ndarraytypes.h and
is only available after including numpy/ndarrayobject.h as it
requires import_array(). This includes PyDataType_FLAGCHK,
PyDataType_REFCHK and NPY_BEGIN_THREADS_DESCR.
The dtype flags on PyArray_Descr must now be accessed through the
PyDataType_FLAGS inline function to be compatible with both 1.x
and 2.x. This function is defined in npy_2_compat.h to allow
backporting. Most or all users should use PyDataType_FLAGCHK which
is available on 1.x and does not require backporting. Cython users
should use Cython 3. Otherwise access will go through Python unless
they use PyDataType_FLAGCHK instead.
(gh-25816)

Datetime functionality exposed in the C API and Cython bindings

The functions NpyDatetime_ConvertDatetime64ToDatetimeStruct,
NpyDatetime_ConvertDatetimeStructToDatetime64,
NpyDatetime_ConvertPyDateTimeToDatetimeStruct,
NpyDatetime_GetDatetimeISO8601StrLen,
NpyDatetime_MakeISO8601Datetime, and
NpyDatetime_ParseISO8601Datetime have been added to the C API to
facilitate converting between strings, Python datetimes, and NumPy
datetimes in external libraries.

(gh-21199)

Const correctness for the generalized ufunc C API

The NumPy C API's functions for constructing generalized ufuncs
(PyUFunc_FromFuncAndData, PyUFunc_FromFuncAndDataAndSignature,
PyUFunc_FromFuncAndDataAndSignatureAndIdentity) take types and
data arguments that are not modified by NumPy's internals. Like the
name and doc arguments, third-party Python extension modules are
likely to supply these arguments from static constants. The types and
data arguments are now const-correct: they are declared as
const char *types and void *const *data, respectively. C code should
not be affected, but C++ code may be.

(gh-23847)

Larger `NPY_MAXDIMS` and `NPY_MAXARGS`, `NPY_RAVEL_AXIS` introduced

NPY_MAXDIMS is now 64, you may want to review its use. This is usually
used in a stack allocation, where the increase should be safe. However,
we do encourage generally to remove any use of NPY_MAXDIMS and
NPY_MAXARGS to eventually allow removing the constraint completely.
For the conversion helper and C-API functions mirroring Python ones such as
take, NPY_MAXDIMS was used to mean axis=None. Such usage must be replaced
with NPY_RAVEL_AXIS. See also migration_maxdims.

(gh-25149)

`NPY_MAXARGS` not constant and `PyArrayMultiIterObject` size change

Since NPY_MAXARGS was increased, it is now a runtime constant and not
compile-time constant anymore. We expect almost no users to notice this.
But if used for stack allocations it now must be replaced with a custom
constant using NPY_MAXARGS as an additional runtime check.

The sizeof(PyArrayMultiIterObject) no longer includes the full size of
the object. We expect nobody to notice this change. It was necessary to
avoid issues with Cython.

(gh-25271)

Required changes for custom legacy user dtypes

In order to improve our DTypes it is unfortunately necessary to break
the ABI, which requires some changes for dtypes registered with
PyArray_RegisterDataType. Please see the documentation of
PyArray_RegisterDataType for how to adapt your code and achieve
compatibility with both 1.x and 2.x.

(gh-25792)

New Public DType API

The C implementation of the NEP 42 DType API is now public. While the
DType API has shipped in NumPy for a few versions, it was only usable in
sessions with a special environment variable set. It is now possible to
write custom DTypes outside of NumPy using the new DType API and the
normal import_array() mechanism for importing the numpy C API.

See dtype-api for more details about the API. As always with a new feature,
please report any bugs you run into implementing or using a new DType. It is
likely that downstream C code that works with dtypes will need to be updated to
work correctly with new DTypes.

(gh-25754)

New C-API import functions

We have now added PyArray_ImportNumPyAPI and PyUFunc_ImportUFuncAPI
as static inline functions to import the NumPy C-API tables. The new
functions have two advantages over import_array and import_ufunc:

They check whether the import was already performed and are
light-weight if not, allowing to add them judiciously (although this
is not preferable in most cases).
The old mechanisms were macros rather than functions which included
a return statement.

The PyArray_ImportNumPyAPI() function is included in npy_2_compat.h
for simpler backporting.

(gh-25866)

Structured dtype information access through functions

The dtype structures fields c_metadata, names, fields, and
subarray must now be accessed through new functions following the same
names, such as PyDataType_NAMES. Direct access of the fields is not
valid as they do not exist for all PyArray_Descr instances. The
metadata field is kept, but the macro version should also be
preferred.

(gh-25802)

Descriptor `elsize` and `alignment` access

Unless compiling only with NumPy 2 support, the elsize and aligment
fields must now be accessed via PyDataType_ELSIZE,
PyDataType_SET_ELSIZE, and PyDataType_ALIGNMENT. In cases where the
descriptor is attached to an array, we advise using PyArray_ITEMSIZE
as it exists on all NumPy versions. Please see
migration_c_descr for more information.

(gh-25943)

NumPy 2.0 C API removals

npy_interrupt.h and the corresponding macros like NPY_SIGINT_ON
have been removed. We recommend querying PyErr_CheckSignals() or
PyOS_InterruptOccurred() periodically (these do currently require
holding the GIL though).
The noprefix.h header has been removed. Replace missing symbols
with their prefixed counterparts (usually an added NPY_ or
npy_).
(gh-23919)
PyUFunc_GetPyVals, PyUFunc_handlefperr, and PyUFunc_checkfperr
have been removed. If needed, a new backwards compatible function to
raise floating point errors could be restored. Reason for removal:
there are no known users and the functions would have made
with np.errstate() fixes much more difficult).
(gh-23922)
The numpy/old_defines.h which was part of the API deprecated since
NumPy 1.7 has been removed. This removes macros of the form
PyArray_CONSTANT. The
replace_old_macros.sed
script may be useful to convert them to the NPY_CONSTANT version.
(gh-24011)
The legacy_inner_loop_selector member of the ufunc struct is
removed to simplify improvements to the dispatching system. There
are no known users overriding or directly accessing this member.
(gh-24271)
NPY_INTPLTR has been removed to avoid confusion (see intp
redefinition).
(gh-24888)
The advanced indexing MapIter and related API has been removed.
The (truly) public part of it was not well tested and had only one
known user (Theano). Making it private will simplify improvements to
speed up ufunc.at, make advanced indexing more maintainable, and
was important for increasing the maximum number of dimensions of
arrays to 64. Please let us know if this API is important to you so
we can find a solution together.
(gh-25138)
The NPY_MAX_ELSIZE macro has been removed, as it only ever
reflected builtin numeric types and served no internal purpose.
(gh-25149)
PyArray_REFCNT and NPY_REFCOUNT are removed. Use Py_REFCNT
instead.
(gh-25156)
PyArrayFlags_Type and PyArray_NewFlagsObject as well as
PyArrayFlagsObject are private now. There is no known use-case;
use the Python API if needed.
PyArray_MoveInto, PyArray_CastTo, PyArray_CastAnyTo are
removed use PyArray_CopyInto and if absolutely needed
PyArray_CopyAnyInto (the latter does a flat copy).
PyArray_FillObjectArray is removed, its only true use was for
implementing np.empty. Create a new empty array or use
PyArray_FillWithScalar() (decrefs existing objects).
PyArray_CompareUCS4 and PyArray_CompareString are removed. Use
the standard C string comparison functions.
PyArray_ISPYTHON is removed as it is misleading, has no known
use-cases, and is easy to replace.
PyArray_FieldNames is removed, as it is unclear what it would be
useful for. It also has incorrect semantics in some possible
use-cases.
PyArray_TypestrConvert is removed, since it seems a misnomer and
unlikely to be used by anyone. If you know the size or are limited
to few types, just use it explicitly, otherwise go via Python
strings.
(gh-25292)
PyDataType_GetDatetimeMetaData is removed, it did not actually do
anything since at least NumPy 1.7.
(gh-25802)
PyArray_GetCastFunc is removed. Note that custom legacy user
dtypes can still provide a castfunc as their implementation, but any
access to them is now removed. The reason for this is that NumPy
never used these internally for many years. If you use simple
numeric types, please just use C casts directly. In case you require
an alternative, please let us know so we can create new API such as
PyArray_CastBuffer() which could use old or new cast functions
depending on the NumPy version.
(gh-25161)

New Features

`np.add` was extended to work with `unicode` and `bytes` dtypes.

(gh-24858)

A new `bitwise_count` function

This new function counts the number of 1-bits in a number.
numpy.bitwise_count works on all the numpy integer types
and integer-like objects.

>>> a = np.array([2**i - 1 for i in range(16)])
>>> np.bitwise_count(a)
array([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15],
      dtype=uint8)

(gh-19355)

macOS Accelerate support, including the ILP64

Support for the updated Accelerate BLAS/LAPACK library, including ILP64
(64-bit integer) support, in macOS 13.3 has been added. This brings
arm64 support, and significant performance improvements of up to 10x for
commonly used linear algebra operations. When Accelerate is selected at
build time, or if no explicit BLAS library selection is done, the 13.3+
version will automatically be used if available.

(gh-24053)

Binary wheels are also available. On macOS >=14.0, users who install
NumPy from PyPI will get wheels built against Accelerate rather than
OpenBLAS.

(gh-25255)

Option to use weights for quantile and percentile functions

A weights keyword is now available for numpy.quantile, numpy.percentile,
numpy.nanquantile and numpy.nanpercentile. Only method="inverted_cdf"
supports weights.

(gh-24254)

Improved CPU optimization tracking

A new tracer mechanism is available which enables tracking of the
enabled targets for each optimized function (i.e., that uses
hardware-specific SIMD instructions) in the NumPy library. With this
enhancement, it becomes possible to precisely monitor the enabled CPU
dispatch targets for the dispatched functions.

A new function named opt_func_info has been added to the new namespace
numpy.lib.introspect, offering this tracing capability. This function allows
you to retrieve information about the enabled targets based on function names
and data type signatures.

(gh-24420)

A new Meson backend for `f2py`

f2py in compile mode (i.e. f2py -c) now accepts the
--backend meson option. This is the default option for Python >=3.12.
For older Python versions, f2py will still default to
--backend distutils.

To support this in realistic use-cases, in compile mode f2py takes a
--dep flag one or many times which maps to dependency() calls in the
meson backend, and does nothing in the distutils backend.

There are no changes for users of f2py only as a code generator, i.e.
without -c.

(gh-24532)

`bind(c)` support for `f2py`

Both functions and subroutines can be annotated with bind(c). f2py
will handle both the correct type mapping, and preserve the unique label
for other C interfaces.

Note: bind(c, name = 'routine_name_other_than_fortran_routine') is
not honored by the f2py bindings by design, since bind(c) with the
name is meant to guarantee only the same name in C and Fortran, not in
Python and Fortran.

(gh-24555)

A new `strict` option for several testing functions

The strict keyword is now available for numpy.testing.assert_allclose,
numpy.testing.assert_equal, and numpy.testing.assert_array_less. Setting
strict=True will disable the broadcasting behaviour for scalars and ensure
that input arrays have the same data type.

(gh-24680,
gh-24770,
gh-24775)

Add `np.core.umath.find` and `np.core.umath.rfind` UFuncs

Add two find and rfind UFuncs that operate on unicode or byte
strings and are used in np.char. They operate similar to str.find
and str.rfind.

(gh-24868)

`diagonal` and `trace` for `numpy.linalg`

numpy.linalg.diagonal and numpy.linalg.trace have been added, which are
array API standard-compatible variants of numpy.diagonal and numpy.trace.
They differ in the default axis selection which define 2-D sub-arrays.

(gh-24887)

New `long` and `ulong` dtypes

numpy.long and numpy.ulong have been added as NumPy integers mapping to
C's long and unsigned long. Prior to NumPy 1.24, numpy.long was an alias
to Python's int.

(gh-24922)

`svdvals` for `numpy.linalg`

numpy.linalg.svdvals has been added. It computes singular values for (a stack
of) matrices. Executing np.svdvals(x) is the same as calling np.svd(x, compute_uv=False, hermitian=False). This function is compatible with the array
API standard.

(gh-24940)

A new `isdtype` function

numpy.isdtype was added to provide a canonical way to classify NumPy's
dtypes in compliance with the array API standard.

(gh-25054)

A new `astype` function

numpy.astype was added to provide an array API standard-compatible
alternative to the numpy.ndarray.astype method.

(gh-25079)

Array API compatible functions' aliases

13 aliases for existing functions were added to improve compatibility
with the array API standard:

Trigonometry: acos, acosh, asin, asinh, atan, atanh,
atan2.
Bitwise: bitwise_left_shift, bitwise_invert,
bitwise_right_shift.
Misc: concat, permute_dims, pow.
In numpy.linalg: tensordot, matmul.

(gh-25086)

New `unique_*` functions

The numpy.unique_all, numpy.unique_counts, numpy.unique_inverse, and
numpy.unique_values functions have been added. They provide functionality of
numpy.unique with different sets of flags. They are array API
standard-compatible, and because the number of arrays they return does not
depend on the values of input arguments, they are easier to target for JIT
compilation.

(gh-25088)

Matrix transpose support for ndarrays

NumPy now offers support for calculating the matrix transpose of an
array (or stack of arrays). The matrix transpose is equivalent to
swapping the last two axes of an array. Both np.ndarray and
np.ma.MaskedArray now expose a .mT attribute, and there is a
matching new numpy.matrix_transpose function.

(gh-23762)

Array API compatible functions for `numpy.linalg`

Six new functions and two aliases were added to improve compatibility
with the Array API standard for `numpy.linalg`:

numpy.linalg.matrix_norm - Computes the matrix norm of
a matrix (or a stack of matrices).
numpy.linalg.vector_norm - Computes the vector norm of
a vector (or batch of vectors).
numpy.vecdot - Computes the (vector) dot product of
two arrays.
numpy.linalg.vecdot - An alias for
numpy.vecdot.
numpy.linalg.matrix_transpose - An alias for
numpy.matrix_transpose.
(gh-25155)
numpy.linalg.outer has been added. It computes the
outer product of two vectors. It differs from
numpy.outer by accepting one-dimensional arrays only.
This function is compatible with the array API standard.
(gh-25101)
numpy.linalg.cross has been added. It computes the
cross product of two (arrays of) 3-dimensional vectors. It differs
from numpy.cross by accepting three-dimensional
vectors only. This function is compatible with the array API
standard.
(gh-25145)

A `correction` argument for `var` and `std`

A correction argument was added to numpy.var and numpy.std, which is an
array API standard compatible alternative to ddof. As both arguments serve a
similar purpose, only one of them can be provided at the same time.

(gh-25169)

`ndarray.device` and `ndarray.to_device`

An ndarray.device attribute and ndarray.to_device method were added
to numpy.ndarray for array API standard compatibility.

Additionally, device keyword-only arguments were added to:
numpy.asarray, numpy.arange, numpy.empty, numpy.empty_like,
numpy.eye, numpy.full, numpy.full_like, numpy.linspace, numpy.ones,
numpy.ones_like, numpy.zeros, and numpy.zeros_like.

For all these new arguments, only device="cpu" is supported.

(gh-25233)

StringDType has been added to NumPy

We have added a new variable-width UTF-8 encoded string data type, implementing
a "NumPy array of Python strings", including support for a user-provided
missing data sentinel. It is intended as a drop-in replacement for arrays of
Python strings and missing data sentinels using the object dtype. See
NEP 55 and the documentation
of stringdtype for more details.

(gh-25347)

New keywords for `cholesky` and `pinv`

The upper and rtol keywords were added to
numpy.linalg.cholesky and numpy.linalg.pinv,
respectively, to improve array API standard compatibility.

For numpy.linalg.pinv, if neither rcond nor rtol is
specified, the rcond's default is used. We plan to deprecate and
remove rcond in the future.

(gh-25388)

New keywords for `sort`, `argsort` and `linalg.matrix_rank`

New keyword parameters were added to improve array API standard
compatibility:

rtol was added to numpy.linalg.matrix_rank.
stable was added to numpy.sort and
numpy.argsort.

(gh-25437)

New `numpy.strings` namespace for string ufuncs

NumPy now implements some string operations as ufuncs. The old np.char
namespace is still available, and where possible the string manipulation
functions in that namespace have been updated to use the new ufuncs,
substantially improving their performance.

Where possible, we suggest updating code to use functions in
np.strings instead of np.char. In the future we may deprecate
np.char in favor of np.strings.

(gh-25463)

`numpy.fft` support for different precisions and in-place calculations

The various FFT routines in numpy.fft now do their
calculations natively in float, double, or long double precision,
depending on the input precision, instead of always calculating in
double precision. Hence, the calculation will now be less precise for
single and more precise for long double precision. The data type of the
output array will now be adjusted accordingly.

Furthermore, all FFT routines have gained an out argument that can be
used for in-place calculations.

(gh-25536)

configtool and pkg-config support

A new numpy-config CLI script is available that can be queried for the
NumPy version and for compile flags needed to use the NumPy C API. This
will allow build systems to better support the use of NumPy as a
dependency. Also, a numpy.pc pkg-config file is now included with
Numpy. In order to find its location for use with PKG_CONFIG_PATH, use
numpy-config --pkgconfigdir.

(gh-25730)

Array API standard support in the main namespace

The main numpy namespace now supports the array API standard. See
array-api-standard-compatibility for
details.

(gh-25911)

Improvements

Strings are now supported by `any`, `all`, and the logical ufuncs.

(gh-25651)

Integer sequences as the shape argument for `memmap`

numpy.memmap can now be created with any integer sequence
as the shape argument, such as a list or numpy array of integers.
Previously, only the types of tuple and int could be used without
raising an error.

(gh-23729)

`errstate` is now faster and context safe

The numpy.errstate context manager/decorator is now faster
and safer. Previously, it was not context safe and had (rare) issues
with thread-safety.

(gh-23936)

AArch64 quicksort speed improved by using Highway's VQSort

The first introduction of the Google Highway library, using VQSort on
AArch64. Execution time is improved by up to 16x in some cases, see the
PR for benchmark results. Extensions to other platforms will be done in
the future.

(gh-24018)

Complex types - underlying C type changes

The underlying C types for all of NumPy's complex types have been
changed to use C99 complex types.
While this change does not affect the memory layout of complex
types, it changes the API to be used to directly retrieve or write
the real or complex part of the complex number, since direct field
access (as in c.real or c.imag) is no longer an option. You can
now use utilities provided in numpy/npy_math.h to do these
operations, like this:
```
npy_cdouble c;
npy_csetreal(&c, 1.0);
npy_csetimag(&c, 0.0);
printf("%d + %di\n", npy_creal(c), npy_cimag(c));
```
To ease cross-version compatibility, equivalent macros and a
compatibility layer have been added which can be used by downstream
packages to continue to support both NumPy 1.x and 2.x. See
complex-numbers for more info.
numpy/npy_common.h now includes complex.h, which means that
complex is now a reserved keyword.

(gh-24085)

`iso_c_binding` support and improved common blocks for `f2py`

Previously, users would have to define their own custom f2cmap file to
use type mappings defined by the Fortran2003 iso_c_binding intrinsic
module. These type maps are now natively supported by f2py

(gh-24555)

f2py now handles common blocks which have kind specifications from
modules. This further expands the usability of intrinsics like
iso_fortran_env and iso_c_binding.

(gh-25186)

Call `str` automatically on third argument to functions like `assert_equal`

The third argument to functions like
numpy.testing.assert_equal now has str called on it
automatically. This way it mimics the built-in assert statement, where
assert_equal(a, b, obj) works like assert a == b, obj.

(gh-24877)

Support for array-like `atol`/`rtol` in `isclose`, `allclose`

The keywords atol and rtol in numpy.isclose and
numpy.allclose now accept both scalars and arrays. An
array, if given, must broadcast to the shapes of the first two array
arguments.

(gh-24878)

Consistent failure messages in test functions

Previously, some numpy.testing assertions printed messages
that referred to the actual and desired results as x and y. Now,
these values are consistently referred to as ACTUAL and DESIRED.

(gh-24931)

n-D FFT transforms allow `s[i] == -1`

The numpy.fft.fftn, numpy.fft.ifftn,
numpy.fft.rfftn, numpy.fft.irfftn,
numpy.fft.fft2, numpy.fft.ifft2,
numpy.fft.rfft2 and numpy.fft.irfft2
functions now use the whole input array along the axis i if
s[i] == -1, in line with the array API standard.

(gh-25495)

Guard PyArrayScalar_VAL and PyUnicodeScalarObject for the limited API

PyUnicodeScalarObject holds a PyUnicodeObject, which is not
available when using Py_LIMITED_API. Add guards to hide it and
consequently also make the PyArrayScalar_VAL macro hidden.

(gh-25531)

Changes

np.gradient() now returns a tuple rather than a list making the
return value immutable.
(gh-23861)
Being fully context and thread-safe, np.errstate can only be
entered once now.
np.setbufsize is now tied to np.errstate(): leaving an
np.errstate context will also reset the bufsize.
(gh-23936)
A new public np.lib.array_utils submodule has been introduced and
it currently contains three functions: byte_bounds (moved from
np.lib.utils), normalize_axis_tuple and normalize_axis_index.
(gh-24540)
Introduce numpy.bool as the new canonical name for
NumPy's boolean dtype, and make numpy.bool\_ an alias
to it. Note that until NumPy 1.24, np.bool was an alias to
Python's builtin bool. The new name helps with array API standard
compatibility and is a more intuitive name.
(gh-25080)
The dtype.flags value was previously stored as a signed integer.
This means that the aligned dtype struct flag lead to negative flags
being set (-128 rather than 128). This flag is now stored unsigned
(positive). Code which checks flags manually may need to adapt. This
may include code compiled with Cython 0.29.x.
(gh-25816)

Representation of NumPy scalars changed

As per NEP 51, the scalar representation has been updated to include the type
information to avoid confusion with Python scalars.

Scalars are now printed as np.float64(3.0) rather than just 3.0.
This may disrupt workflows that store representations of numbers (e.g.,
to files) making it harder to read them. They should be stored as
explicit strings, for example by using str() or f"{scalar!s}". For
the time being, affected users can use
np.set_printoptions(legacy="1.25") to get the old behavior (with
possibly a few exceptions). Documentation of downstream projects may
require larger updates, if code snippets are tested. We are working on
tooling for
doctest-plus
to facilitate updates.

(gh-22449)

Truthiness of NumPy strings changed

NumPy strings previously were inconsistent about how they defined if the
string is True or False and the definition did not match the one
used by Python. Strings are now considered True when they are
non-empty and False when they are empty. This changes the following
distinct cases:

Casts from string to boolean were previously roughly equivalent to
string_array.astype(np.int64).astype(bool), meaning that only
valid integers could be cast. Now a string of "0" will be
considered True since it is not empty. If you need the old
behavior, you may use the above step (casting to integer first) or
string_array == "0" (if the input is only ever 0 or 1). To get
the new result on old NumPy versions use string_array != "".
np.nonzero(string_array) previously ignored whitespace so that a
string only containing whitespace was considered False. Whitespace
is now considered True.

This change does not affect np.loadtxt, np.fromstring, or
np.genfromtxt. The first two still use the integer definition, while
genfromtxt continues to match for "true" (ignoring case). However,
if np.bool_ is used as a converter the result will change.

The change does affect np.fromregex as it uses direct assignments.

(gh-23871)

A `mean` keyword was added to var and std function

Often when the standard deviation is needed the mean is also needed. The
same holds for the variance and the mean. Until now the mean is then
calculated twice, the change introduced here for the numpy.var and
numpy.std functions allows for passing in a precalculated mean as an keyword
argument. See the docstrings for details and an example illustrating the
speed-up.

(gh-24126)

Remove datetime64 deprecation warning when constructing with timezone

The numpy.datetime64 method now issues a UserWarning rather than a
DeprecationWarning whenever a timezone is included in the datetime string that
is provided.

(gh-24193)

Default integer dtype is now 64-bit on 64-bit Windows

The default NumPy integer is now 64-bit on all 64-bit systems as the
historic 32-bit default on Windows was a common source of issues. Most
users should not notice this. The main issues may occur with code
interfacing with libraries written in a compiled language like C. For
more information see migration_windows_int64.

(gh-24224)

Renamed `numpy.core` to `numpy._core`

Accessing numpy.core now emits a DeprecationWarning. In practice we
have found that most downstream usage of numpy.core was to access
functionality that is available in the main numpy namespace. If for
some reason you are using functionality in numpy.core that is not
available in the main numpy namespace, this means you are likely using
private NumPy internals. You can still access these internals via
numpy._core without a deprecation warning but we do not provide any
backward compatibility guarantees for NumPy internals. Please open an
issue if you think a mistake was made and something needs to be made
public.

(gh-24634)

The "relaxed strides" debug build option, which was previously enabled
through the NPY_RELAXED_STRIDES_DEBUG environment variable or the
-Drelaxed-strides-debug config-settings flag has been removed.

(gh-24717)

Redefinition of `np.intp`/`np.uintp` (almost never a change)

Due to the actual use of these types almost always matching the use of
size_t/Py_ssize_t this is now the definition in C. Previously, it
matched intptr_t and uintptr_t which would often have been subtly
incorrect. This has no effect on the vast majority of machines since the
size of these types only differ on extremely niche platforms.

However, it means that:

Pointers may not necessarily fit into an intp typed array anymore.
The p and P character codes can still be used, however.
Creating intptr_t or uintptr_t typed arrays in C remains
possible in a cross-platform way via PyArray_DescrFromType('p').
The new character codes nN were introduced.
It is now correct to use the Python C-API functions when parsing to
npy_intp typed arguments.

(gh-24888)

`numpy.fft.helper` made private

numpy.fft.helper was renamed to numpy.fft._helper to indicate that
it is a private submodule. All public functions exported by it should be
accessed from numpy.fft.

(gh-24945)

`numpy.linalg.linalg` made private

numpy.linalg.linalg was renamed to numpy.linalg._linalg to indicate
that it is a private submodule. All public functions exported by it
should be accessed from numpy.linalg.

(gh-24946)

Out-of-bound axis not the same as `axis=None`

In some cases axis=32 or for concatenate any large value was the same
as axis=None. Except for concatenate this was deprecate. Any out of
bound axis value will now error, make sure to use axis=None.

(gh-25149)

New `copy` keyword meaning for `array` and `asarray` constructors

Now numpy.array and numpy.asarray support
three values for copy parameter:

None - A copy will only be made if it is necessary.
True - Always make a copy.
False - Never make a copy. If a copy is required a ValueError is
raised.

The meaning of False changed as it now raises an exception if a copy
is needed.

(gh-25168)

The `array` special method now takes a `copy` keyword argument.

NumPy will pass copy to the __array__ special method in situations
where it would be set to a non-default value (e.g. in a call to
np.asarray(some_object, copy=False)). Currently, if an unexpected
keyword argument error is raised after this, NumPy will print a warning
and re-try without the copy keyword argument. Implementations of
objects implementing the __array__ protocol should accept a copy
keyword argument with the same meaning as when passed to
numpy.array or numpy.asarray.

(gh-25168)

Cleanup of initialization of `numpy.dtype` with strings with commas

The interpretation of strings with commas is changed slightly, in that a
trailing comma will now always create a structured dtype. E.g., where
previously np.dtype("i") and np.dtype("i,") were treated as
identical, now np.dtype("i,") will create a structured dtype, with a
single field. This is analogous to np.dtype("i,i") creating a
structured dtype with two fields, and makes the behaviour consistent
with that expected of tuples.

At the same time, the use of single number surrounded by parenthesis to
indicate a sub-array shape, like in np.dtype("(2)i,"), is deprecated.
Instead; one should use np.dtype("(2,)i") or np.dtype("2i").
Eventually, using a number in parentheses will raise an exception, like
is the case for initializations without a comma, like
np.dtype("(2)i").

(gh-25434)

Change in how complex sign is calculated

Following the array API standard, the complex sign is now calculated as
z / |z| (instead of the rather less logical case where the sign of the
real part was taken, unless the real part was zero, in which case the
sign of the imaginary part was returned). Like for real numbers, zero is
returned if z==0.

(gh-25441)

Return types of functions that returned a list of arrays

Functions that returned a list of ndarrays have been changed to return a
tuple of ndarrays instead. Returning tuples consistently whenever a
sequence of arrays is returned makes it easier for JIT compilers like
Numba, as well as for static type checkers in some cases, to support
these functions. Changed functions are: numpy.atleast_1d, numpy.atleast_2d,
numpy.atleast_3d, numpy.broadcast_arrays, numpy.meshgrid,
numpy.ogrid, numpy.histogramdd.

`np.unique` `return_inverse` shape for multi-dimensional inputs

When multi-dimensional inputs are passed to np.unique with
return_inverse=True, the unique_inverse output is now shaped such
that the input can be reconstructed directly using
np.take(unique, unique_inverse) when axis=None, and
np.take_along_axis(unique, unique_inverse, axis=axis) otherwise.

(gh-25553,
gh-25570)

`any` and `all` return booleans for object arrays

The any and all functions and methods now return booleans also for
object arrays. Previously, they did a reduction which behaved like the
Python or and and operators which evaluates to one of the arguments.
You can use np.logical_or.reduce and np.logical_and.reduce to
achieve the previous behavior.

(gh-25712)

Content from release note snippets in doc/release/upcoming_changes:

Checksums

MD5

b2f97f907cc640f5f619ea4ebd1231d3  numpy-2.0.0b1-cp310-cp310-macosx_10_9_x86_64.whl
db158043b6fad6e523e23b3eb2de5d88  numpy-2.0.0b1-cp310-cp310-macosx_11_0_arm64.whl
39086961c062d97c5b42da057b9b1947  numpy-2.0.0b1-cp310-cp310-macosx_14_0_arm64.whl
3362d35bf69b852b98b41b8373253a0f  numpy-2.0.0b1-cp310-cp310-macosx_14_0_x86_64.whl
66e907969e32ec43e887cabcc1884763  numpy-2.0.0b1-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
b8d1bece144e3b6aae641d44821f815f  numpy-2.0.0b1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
96ab156ec312bb451e8c5e19de4a28b7  numpy-2.0.0b1-cp310-cp310-musllinux_1_1_aarch64.whl
c04819a4f3395b81d124ffc6330925e9  numpy-2.0.0b1-cp310-cp310-musllinux_1_1_x86_64.whl
6af68b8eb8fe583ffabab9bd7da1c620  numpy-2.0.0b1-cp310-cp310-win32.whl
3b8a9514e5795985bcba20e213d55b54  numpy-2.0.0b1-cp310-cp310-win_amd64.whl
0128ad9249f70d97a057a23e0cef1515  numpy-2.0.0b1-cp311-cp311-macosx_10_9_x86_64.whl
612c018a7676ce3747cb863762750e1d  numpy-2.0.0b1-cp311-cp311-macosx_11_0_arm64.whl
6b1480446aff53c71c903fc1248bca94  numpy-2.0.0b1-cp311-cp311-macosx_14_0_arm64.whl
8d66a0af99edf30dc9de487b3f8c1639  numpy-2.0.0b1-cp311-cp311-macosx_14_0_x86_64.whl
f9154a0885b2647d7e81f32900390ebb  numpy-2.0.0b1-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
9dd14e2b594a2d47eb25ecc759d5adaa  numpy-2.0.0b1-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
8434d07fc4eb80c5df9ae5ebf95546eb  numpy-2.0.0b1-cp311-cp311-musllinux_1_1_aarch64.whl
a0402697c93a9d6bc8d979fabd6bf179  numpy-2.0.0b1-cp311-cp311-musllinux_1_1_x86_64.whl
2ba67ffb4b92b54394b6929b3a899cb2  numpy-2.0.0b1-cp311-cp311-win32.whl
d75e2f02c698e492b7b07f0659f9bbe4  numpy-2.0.0b1-cp311-cp311-win_amd64.whl
558fefd135de6fcebe2b94d857a84c32  numpy-2.0.0b1-cp312-cp312-macosx_10_9_x86_64.whl
d684790e4509e7daa99a1aef1d0be536  numpy-2.0.0b1-cp312-cp312-macosx_11_0_arm64.whl
fd5d4f1d1da0cc685c54e9abd2f9dceb  numpy-2.0.0b1-cp312-cp312-macosx_14_0_arm64.whl
65183c1302348d3db60eaf3b62c1e577  numpy-2.0.0b1-cp312-cp312-macosx_14_0_x86_64.whl
305eaf68e214011557303988f4635271  numpy-2.0.0b1-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
e3a84e27effd888cf93eb2c1aad759e7  numpy-2.0.0b1-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
88eef47ecdd11ac0939291abe0c74b6f  numpy-2.0.0b1-cp312-cp312-musllinux_1_1_aarch64.whl
fd390078c0046c20a659035c1826185f  numpy-2.0.0b1-cp312-cp312-musllinux_1_1_x86_64.whl
23db11989d2d0086ff12655355245a2a  numpy-2.0.0b1-cp312-cp312-win32.whl
323d05ef29a9c8166d865ab221faf7dc  numpy-2.0.0b1-cp312-cp312-win_amd64.whl
f5ad7adf599b65050ccd116802f0265d  numpy-2.0.0b1-cp39-cp39-macosx_10_9_x86_64.whl
89a94dddb18e4210e01ee6ca24012fcb  numpy-2.0.0b1-cp39-cp39-macosx_11_0_arm64.whl
409a537dc5ea249b3e6868dd37932342  numpy-2.0.0b1-cp39-cp39-macosx_14_0_arm64.whl
0db893de846425d58b90f05c1db3d191  numpy-2.0.0b1-cp39-cp39-macosx_14_0_x86_64.whl
c73ba41d166a5f2e72cdc48b8554c6e6  numpy-2.0.0b1-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
786236fc9099283255133273535b8de0  numpy-2.0.0b1-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
f2e8717957a6b3b37f881e8939a2af37  numpy-2.0.0b1-cp39-cp39-musllinux_1_1_aarch64.whl
dad671b45f6e13c28ead06064b03eaee  numpy-2.0.0b1-cp39-cp39-musllinux_1_1_x86_64.whl
76f8f89ff91d06df684cf47d7ea6d8ab  numpy-2.0.0b1-cp39-cp39-win32.whl
d4dcbd6157783aa0e78710549f13876f  numpy-2.0.0b1-cp39-cp39-win_amd64.whl
41a13de3afff77390b0d1ea3c7e407db  numpy-2.0.0b1-pp39-pypy39_pp73-macosx_10_9_x86_64.whl
fc2ff82233376f853161c7f9bc6d44b7  numpy-2.0.0b1-pp39-pypy39_pp73-macosx_14_0_arm64.whl
860609ee9f1f24d4f28fbbcf3d31cdc9  numpy-2.0.0b1-pp39-pypy39_pp73-macosx_14_0_x86_64.whl
2a97175cec7a5b1280ed2a991fea23ff  numpy-2.0.0b1-pp39-pypy39_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
1656013175e650e053c15fd886be58f1  numpy-2.0.0b1-pp39-pypy39_pp73-win_amd64.whl
c06e95d7cadfa33a1f4549c9a5dcba05  numpy-2.0.0b1.tar.gz

SHA256

411ed8eb48eb679fc732f22e90c9adb994ec6ad2d9c2f53593325a975f9fa501  numpy-2.0.0b1-cp310-cp310-macosx_10_9_x86_64.whl
f8aca0561166702070ea9abcafd70da44df48be70d16f0a886e359127436fdcc  numpy-2.0.0b1-cp310-cp310-macosx_11_0_arm64.whl
0d217dae0f20a3400c1d80aa8401af9de93b9bb4ea7518b8ba200ff8ff62529e  numpy-2.0.0b1-cp310-cp310-macosx_14_0_arm64.whl
824351cb4cce66c1f8e16c1698c01de8d5e4197461f78197c327281f107fc1b2  numpy-2.0.0b1-cp310-cp310-macosx_14_0_x86_64.whl
cae0959a4f5a9c16896a87a43c9e81384f48b69f835f55050948071488820486  numpy-2.0.0b1-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
3d47a42c1e48e46dbbe32e0395f8aa6e8ddd251771ed9ec47fc07aa89b8aac89  numpy-2.0.0b1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
909024f923c019d1b9dca16871844f1c313c422bd430a0b7e4a24a3acb766483  numpy-2.0.0b1-cp310-cp310-musllinux_1_1_aarch64.whl
fc6e82bea99727aeed964808f26bed95323825a75e94c015eb913fb6ec3dbdf8  numpy-2.0.0b1-cp310-cp310-musllinux_1_1_x86_64.whl
36862cad55650afbcb3f0e3a5edc07ba4c1090eb649208a41fadcf82cf1b2966  numpy-2.0.0b1-cp310-cp310-win32.whl
0e6a63c725143a6be0e48effcf01b8361b80ab20e2444704356f9d9db48ba429  numpy-2.0.0b1-cp310-cp310-win_amd64.whl
e6c3ba4bcb6cf3fd4ace244075fa214b4f0c090f12437378200a2de68144c166  numpy-2.0.0b1-cp311-cp311-macosx_10_9_x86_64.whl
89bbb14534e53c6175aabc8449a8bdf83f02da62f13d1b5facbb2fd1fecae2e2  numpy-2.0.0b1-cp311-cp311-macosx_11_0_arm64.whl
b14b6e6ca51afdcfc589cb9d6fb73aedf38009a1a0ecab15f77e3d0e0754cac0  numpy-2.0.0b1-cp311-cp311-macosx_14_0_arm64.whl
ffef68423c1edc5d10321f9787fb9d8c20a36fc08ffdba863d103924d02dadce  numpy-2.0.0b1-cp311-cp311-macosx_14_0_x86_64.whl
7e8725313b8a8aaa9cfac450713b1a74a8d79ae010ee0d0dd97505abf54d247b  numpy-2.0.0b1-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
d83e18f1c4164dbcaa01adc8f4a3aebc3c5fa635d2009d8dc1bf53dd7eab0063  numpy-2.0.0b1-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
91e37a5bb38c11bde547aefeb79dd382b5d9d1d140931927bca46c9d198e08f3  numpy-2.0.0b1-cp311-cp311-musllinux_1_1_aarch64.whl
d39f1005a627c5960f67b02c1f76f265e0d4219b6d7948a7809dc14443fcbeb6  numpy-2.0.0b1-cp311-cp311-musllinux_1_1_x86_64.whl
3ed4afbcdb8db622b90ef33bf0c0d080f287ec590032f9033be5cbc51e005b66  numpy-2.0.0b1-cp311-cp311-win32.whl
941382abe21d26222310275a91f053386450b5364f1307641d03babfec5b1931  numpy-2.0.0b1-cp311-cp311-win_amd64.whl
a78a38ff86aa651534979d597fdb178c7ae2c9934d95bcc921971ceea14ef54a  numpy-2.0.0b1-cp312-cp312-macosx_10_9_x86_64.whl
e5222fb05011c310d294c40e2b8640c9351aaf3238c0605486a3f041a7befabd  numpy-2.0.0b1-cp312-cp312-macosx_11_0_arm64.whl
0f69c008a8533879ea0480fe11b28154c0dc12567522406f2c887bc549a98865  numpy-2.0.0b1-cp312-cp312-macosx_14_0_arm64.whl
a5b47099876fceefb5ac4d2cfe4ee7337de22253aafe6f2e545b84d100bf9e22  numpy-2.0.0b1-cp312-cp312-macosx_14_0_x86_64.whl
a81816e4dc75351dd1ce2d84f381856b8962eef1757ddfe13007d2a8bb966fda  numpy-2.0.0b1-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
3ecb219af16b0dbf58bbe1fdb4d074582f9a99567d85c630cf82c3b40168a15d  numpy-2.0.0b1-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
7479d8f43dd78a3bd1c8a3c7c9b06e71639c480a0223c31a4aeb2c7e8fd62151  numpy-2.0.0b1-cp312-cp312-musllinux_1_1_aarch64.whl
1665b832541449c7079ee9d41f334ab832a1d84511cc834c0bc8d98bf96d1df5  numpy-2.0.0b1-cp312-cp312-musllinux_1_1_x86_64.whl
585471edf1f205fb589632581cc7b30c6c0e78d79b3c754739bb62ff568fa587  numpy-2.0.0b1-cp312-cp312-win32.whl
881df25d857873947d54dbed01d98c417f3feb5df86ece719eebf1edbbb2095c  numpy-2.0.0b1-cp312-cp312-win_amd64.whl
797dc478feed31f78bca1c69d9a167c6294599927c184f4e9b569ad8895ca6e5  numpy-2.0.0b1-cp39-cp39-macosx_10_9_x86_64.whl
49cb06682f4588c2553a63445b7e37aec731452fe380c3bd142377783a9ba014  numpy-2.0.0b1-cp39-cp39-macosx_11_0_arm64.whl
5fd7ec50b9650ac0aa4fd318eceb9059ed3c0ab3aa79d5f260a10158521f9770  numpy-2.0.0b1-cp39-cp39-macosx_14_0_arm64.whl
72526252a5d1da5067181bfd3df9cc6d7dcd024b757f5d35e8f1d0c08cb729c1  numpy-2.0.0b1-cp39-cp39-macosx_14_0_x86_64.whl
7870b854823217f34e6258328f46e40f68784f61408deb37a29ca64762c60c10  numpy-2.0.0b1-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
26a0978595ac2e8160d27f7537ff94402eaaf3ea7a768e7f99170ed91453d1bf  numpy-2.0.0b1-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
5801a93e424c12366d8b0b411dfeb7102f7429f0934059a39b1529f02ea2606b  numpy-2.0.0b1-cp39-cp39-musllinux_1_1_aarch64.whl
2f67038ecdf4b372d81fa00530547a5d04b77da5b1e4fc55f58021f3135331ea  numpy-2.0.0b1-cp39-cp39-musllinux_1_1_x86_64.whl
f32b6ec16518b3ba1a2d3a100d9b413cf24aaeeefdec19f1cddec55cb4a31dac  numpy-2.0.0b1-cp39-cp39-win32.whl
70a22408ed088725fe44a6f55a077d1f704977b262e53d30ba485a01229028a3  numpy-2.0.0b1-cp39-cp39-win_amd64.whl
a69f1624d036953f3f2795f22e6be452ee6d24937ae14f77c2e536589e20caa2  numpy-2.0.0b1-pp39-pypy39_pp73-macosx_10_9_x86_64.whl
c27970540ee6b4b8325779cd22eee0283cb9dc6511130ff54e774fcd0a261d4b  numpy-2.0.0b1-pp39-pypy39_pp73-macosx_14_0_arm64.whl
393adcc241ff7010b43e4660710a43c322189ff67461afba18bbaf9f5581b221  numpy-2.0.0b1-pp39-pypy39_pp73-macosx_14_0_x86_64.whl
4205b3efa27b74cb096443bdda178f5032ffc6b41306a7d4a0b903b4b614b146  numpy-2.0.0b1-pp39-pypy39_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
52f9cd632f9f5e179e98769d76702ce9a307439f36191607d5ee06cb8a986d01  numpy-2.0.0b1-pp39-pypy39_pp73-win_amd64.whl
e0bb33a37d0d0b9a19cd41a093877f830e06bd4d989341b9792896cf08e629f7  numpy-2.0.0b1.tar.gz

numpy 2.0.0b1 on Python PyPI

NumPy 2.0.0 Release Notes

Highlights

NumPy 2.0 Python API removals

__array_prepare__ is removed

Deprecations

numpy.fft deprecations for n-D transforms with None values in arguments

np.linalg.lstsq now defaults to a new rcond value

Expired deprecations

Compatibility notes

loadtxt and genfromtxt default encoding changed

f2py compatibility notes

Minor changes in behavior of sorting functions

Removed ambiguity when broadcasting in np.solve

Modified representation for Polynomial

C API changes

Datetime functionality exposed in the C API and Cython bindings

Const correctness for the generalized ufunc C API

Larger NPY_MAXDIMS and NPY_MAXARGS, NPY_RAVEL_AXIS introduced

NPY_MAXARGS not constant and PyArrayMultiIterObject size change

Required changes for custom legacy user dtypes

New Public DType API

New C-API import functions

Structured dtype information access through functions

Descriptor elsize and alignment access

NumPy 2.0 C API removals

New Features

np.add was extended to work with unicode and bytes dtypes.

A new bitwise_count function

macOS Accelerate support, including the ILP64

Option to use weights for quantile and percentile functions

Improved CPU optimization tracking

A new Meson backend for f2py

bind(c) support for f2py

A new strict option for several testing functions

Add np.core.umath.find and np.core.umath.rfind UFuncs

diagonal and trace for numpy.linalg

New long and ulong dtypes

svdvals for numpy.linalg

A new isdtype function

A new astype function