Add files using upload-large-folder tool

lib/python3.12/site-packages/_multiprocess/__init__.py

@@ -0,0 +1,8 @@
+#!/usr/bin/env python
+#
+# Author: Mike McKerns (mmckerns @caltech and @uqfoundation)
+# Copyright (c) 2022-2024 The Uncertainty Quantification Foundation.
+# License: 3-clause BSD.  The full license text is available at:
+#  - https://github.com/uqfoundation/multiprocess/blob/master/LICENSE
+
+from _multiprocessing import *

lib/python3.12/site-packages/blinker/__init__.py

@@ -0,0 +1,17 @@
+from __future__ import annotations
+
+from .base import ANY
+from .base import default_namespace
+from .base import NamedSignal
+from .base import Namespace
+from .base import Signal
+from .base import signal
+
+__all__ = [
+    "ANY",
+    "default_namespace",
+    "NamedSignal",
+    "Namespace",
+    "Signal",
+    "signal",
+]

lib/python3.12/site-packages/blinker/_utilities.py

@@ -0,0 +1,64 @@
+from __future__ import annotations
+
+import collections.abc as c
+import inspect
+import typing as t
+from weakref import ref
+from weakref import WeakMethod
+
+T = t.TypeVar("T")
+
+
+class Symbol:
+    """A constant symbol, nicer than ``object()``. Repeated calls return the
+    same instance.
+
+    >>> Symbol('foo') is Symbol('foo')
+    True
+    >>> Symbol('foo')
+    foo
+    """
+
+    symbols: t.ClassVar[dict[str, Symbol]] = {}
+
+    def __new__(cls, name: str) -> Symbol:
+        if name in cls.symbols:
+            return cls.symbols[name]
+
+        obj = super().__new__(cls)
+        cls.symbols[name] = obj
+        return obj
+
+    def __init__(self, name: str) -> None:
+        self.name = name
+
+    def __repr__(self) -> str:
+        return self.name
+
+    def __getnewargs__(self) -> tuple[t.Any, ...]:
+        return (self.name,)
+
+
+def make_id(obj: object) -> c.Hashable:
+    """Get a stable identifier for a receiver or sender, to be used as a dict
+    key or in a set.
+    """
+    if inspect.ismethod(obj):
+        # The id of a bound method is not stable, but the id of the unbound
+        # function and instance are.
+        return id(obj.__func__), id(obj.__self__)
+
+    if isinstance(obj, (str, int)):
+        # Instances with the same value always compare equal and have the same
+        # hash, even if the id may change.
+        return obj
+
+    # Assume other types are not hashable but will always be the same instance.
+    return id(obj)
+
+
+def make_ref(obj: T, callback: c.Callable[[ref[T]], None] | None = None) -> ref[T]:
+    if inspect.ismethod(obj):
+        return WeakMethod(obj, callback)  # type: ignore[arg-type, return-value]
+
+    return ref(obj, callback)

lib/python3.12/site-packages/blinker/base.py

@@ -0,0 +1,512 @@
+from __future__ import annotations
+
+import collections.abc as c
+import sys
+import typing as t
+import weakref
+from collections import defaultdict
+from contextlib import contextmanager
+from functools import cached_property
+from inspect import iscoroutinefunction
+
+from ._utilities import make_id
+from ._utilities import make_ref
+from ._utilities import Symbol
+
+F = t.TypeVar("F", bound=c.Callable[..., t.Any])
+
+ANY = Symbol("ANY")
+"""Symbol for "any sender"."""
+
+ANY_ID = 0
+
+
+class Signal:
+    """A notification emitter.
+
+    :param doc: The docstring for the signal.
+    """
+
+    ANY = ANY
+    """An alias for the :data:`~blinker.ANY` sender symbol."""
+
+    set_class: type[set[t.Any]] = set
+    """The set class to use for tracking connected receivers and senders.
+    Python's ``set`` is unordered. If receivers must be dispatched in the order
+    they were connected, an ordered set implementation can be used.
+
+    .. versionadded:: 1.7
+    """
+
+    @cached_property
+    def receiver_connected(self) -> Signal:
+        """Emitted at the end of each :meth:`connect` call.
+
+        The signal sender is the signal instance, and the :meth:`connect`
+        arguments are passed through: ``receiver``, ``sender``, and ``weak``.
+
+        .. versionadded:: 1.2
+        """
+        return Signal(doc="Emitted after a receiver connects.")
+
+    @cached_property
+    def receiver_disconnected(self) -> Signal:
+        """Emitted at the end of each :meth:`disconnect` call.
+
+        The sender is the signal instance, and the :meth:`disconnect` arguments
+        are passed through: ``receiver`` and ``sender``.
+
+        This signal is emitted **only** when :meth:`disconnect` is called
+        explicitly. This signal cannot be emitted by an automatic disconnect
+        when a weakly referenced receiver or sender goes out of scope, as the
+        instance is no longer be available to be used as the sender for this
+        signal.
+
+        An alternative approach is available by subscribing to
+        :attr:`receiver_connected` and setting up a custom weakref cleanup
+        callback on weak receivers and senders.
+
+        .. versionadded:: 1.2
+        """
+        return Signal(doc="Emitted after a receiver disconnects.")
+
+    def __init__(self, doc: str | None = None) -> None:
+        if doc:
+            self.__doc__ = doc
+
+        self.receivers: dict[
+            t.Any, weakref.ref[c.Callable[..., t.Any]] | c.Callable[..., t.Any]
+        ] = {}
+        """The map of connected receivers. Useful to quickly check if any
+        receivers are connected to the signal: ``if s.receivers:``. The
+        structure and data is not part of the public API, but checking its
+        boolean value is.
+        """
+
+        self.is_muted: bool = False
+        self._by_receiver: dict[t.Any, set[t.Any]] = defaultdict(self.set_class)
+        self._by_sender: dict[t.Any, set[t.Any]] = defaultdict(self.set_class)
+        self._weak_senders: dict[t.Any, weakref.ref[t.Any]] = {}
+
+    def connect(self, receiver: F, sender: t.Any = ANY, weak: bool = True) -> F:
+        """Connect ``receiver`` to be called when the signal is sent by
+        ``sender``.
+
+        :param receiver: The callable to call when :meth:`send` is called with
+            the given ``sender``, passing ``sender`` as a positional argument
+            along with any extra keyword arguments.
+        :param sender: Any object or :data:`ANY`. ``receiver`` will only be
+            called when :meth:`send` is called with this sender. If ``ANY``, the
+            receiver will be called for any sender. A receiver may be connected
+            to multiple senders by calling :meth:`connect` multiple times.
+        :param weak: Track the receiver with a :mod:`weakref`. The receiver will
+            be automatically disconnected when it is garbage collected. When
+            connecting a receiver defined within a function, set to ``False``,
+            otherwise it will be disconnected when the function scope ends.
+        """
+        receiver_id = make_id(receiver)
+        sender_id = ANY_ID if sender is ANY else make_id(sender)
+
+        if weak:
+            self.receivers[receiver_id] = make_ref(
+                receiver, self._make_cleanup_receiver(receiver_id)
+            )
+        else:
+            self.receivers[receiver_id] = receiver
+
+        self._by_sender[sender_id].add(receiver_id)
+        self._by_receiver[receiver_id].add(sender_id)
+
+        if sender is not ANY and sender_id not in self._weak_senders:
+            # store a cleanup for weakref-able senders
+            try:
+                self._weak_senders[sender_id] = make_ref(
+                    sender, self._make_cleanup_sender(sender_id)
+                )
+            except TypeError:
+                pass
+
+        if "receiver_connected" in self.__dict__ and self.receiver_connected.receivers:
+            try:
+                self.receiver_connected.send(
+                    self, receiver=receiver, sender=sender, weak=weak
+                )
+            except TypeError:
+                # TODO no explanation or test for this
+                self.disconnect(receiver, sender)
+                raise
+
+        return receiver
+
+    def connect_via(self, sender: t.Any, weak: bool = False) -> c.Callable[[F], F]:
+        """Connect the decorated function to be called when the signal is sent
+        by ``sender``.
+
+        The decorated function will be called when :meth:`send` is called with
+        the given ``sender``, passing ``sender`` as a positional argument along
+        with any extra keyword arguments.
+
+        :param sender: Any object or :data:`ANY`. ``receiver`` will only be
+            called when :meth:`send` is called with this sender. If ``ANY``, the
+            receiver will be called for any sender. A receiver may be connected
+            to multiple senders by calling :meth:`connect` multiple times.
+        :param weak: Track the receiver with a :mod:`weakref`. The receiver will
+            be automatically disconnected when it is garbage collected. When
+            connecting a receiver defined within a function, set to ``False``,
+            otherwise it will be disconnected when the function scope ends.=
+
+        .. versionadded:: 1.1
+        """
+
+        def decorator(fn: F) -> F:
+            self.connect(fn, sender, weak)
+            return fn
+
+        return decorator
+
+    @contextmanager
+    def connected_to(
+        self, receiver: c.Callable[..., t.Any], sender: t.Any = ANY
+    ) -> c.Generator[None, None, None]:
+        """A context manager that temporarily connects ``receiver`` to the
+        signal while a ``with`` block executes. When the block exits, the
+        receiver is disconnected. Useful for tests.
+
+        :param receiver: The callable to call when :meth:`send` is called with
+            the given ``sender``, passing ``sender`` as a positional argument
+            along with any extra keyword arguments.
+        :param sender: Any object or :data:`ANY`. ``receiver`` will only be
+            called when :meth:`send` is called with this sender. If ``ANY``, the
+            receiver will be called for any sender.
+
+        .. versionadded:: 1.1
+        """
+        self.connect(receiver, sender=sender, weak=False)
+
+        try:
+            yield None
+        finally:
+            self.disconnect(receiver)
+
+    @contextmanager
+    def muted(self) -> c.Generator[None, None, None]:
+        """A context manager that temporarily disables the signal. No receivers
+        will be called if the signal is sent, until the ``with`` block exits.
+        Useful for tests.
+        """
+        self.is_muted = True
+
+        try:
+            yield None
+        finally:
+            self.is_muted = False
+
+    def send(
+        self,
+        sender: t.Any | None = None,
+        /,
+        *,
+        _async_wrapper: c.Callable[
+            [c.Callable[..., c.Coroutine[t.Any, t.Any, t.Any]]], c.Callable[..., t.Any]
+        ]
+        | None = None,
+        **kwargs: t.Any,
+    ) -> list[tuple[c.Callable[..., t.Any], t.Any]]:
+        """Call all receivers that are connected to the given ``sender``
+        or :data:`ANY`. Each receiver is called with ``sender`` as a positional
+        argument along with any extra keyword arguments. Return a list of
+        ``(receiver, return value)`` tuples.
+
+        The order receivers are called is undefined, but can be influenced by
+        setting :attr:`set_class`.
+
+        If a receiver raises an exception, that exception will propagate up.
+        This makes debugging straightforward, with an assumption that correctly
+        implemented receivers will not raise.
+
+        :param sender: Call receivers connected to this sender, in addition to
+            those connected to :data:`ANY`.
+        :param _async_wrapper: Will be called on any receivers that are async
+            coroutines to turn them into sync callables. For example, could run
+            the receiver with an event loop.
+        :param kwargs: Extra keyword arguments to pass to each receiver.
+
+        .. versionchanged:: 1.7
+            Added the ``_async_wrapper`` argument.
+        """
+        if self.is_muted:
+            return []
+
+        results = []
+
+        for receiver in self.receivers_for(sender):
+            if iscoroutinefunction(receiver):
+                if _async_wrapper is None:
+                    raise RuntimeError("Cannot send to a coroutine function.")
+
+                result = _async_wrapper(receiver)(sender, **kwargs)
+            else:
+                result = receiver(sender, **kwargs)
+
+            results.append((receiver, result))
+
+        return results
+
+    async def send_async(
+        self,
+        sender: t.Any | None = None,
+        /,
+        *,
+        _sync_wrapper: c.Callable[
+            [c.Callable[..., t.Any]], c.Callable[..., c.Coroutine[t.Any, t.Any, t.Any]]
+        ]
+        | None = None,
+        **kwargs: t.Any,
+    ) -> list[tuple[c.Callable[..., t.Any], t.Any]]:
+        """Await all receivers that are connected to the given ``sender``
+        or :data:`ANY`. Each receiver is called with ``sender`` as a positional
+        argument along with any extra keyword arguments. Return a list of
+        ``(receiver, return value)`` tuples.
+
+        The order receivers are called is undefined, but can be influenced by
+        setting :attr:`set_class`.
+
+        If a receiver raises an exception, that exception will propagate up.
+        This makes debugging straightforward, with an assumption that correctly
+        implemented receivers will not raise.
+
+        :param sender: Call receivers connected to this sender, in addition to
+            those connected to :data:`ANY`.
+        :param _sync_wrapper: Will be called on any receivers that are sync
+            callables to turn them into async coroutines. For example,
+            could call the receiver in a thread.
+        :param kwargs: Extra keyword arguments to pass to each receiver.
+
+        .. versionadded:: 1.7
+        """
+        if self.is_muted:
+            return []
+
+        results = []
+
+        for receiver in self.receivers_for(sender):
+            if not iscoroutinefunction(receiver):
+                if _sync_wrapper is None:
+                    raise RuntimeError("Cannot send to a non-coroutine function.")
+
+                result = await _sync_wrapper(receiver)(sender, **kwargs)
+            else:
+                result = await receiver(sender, **kwargs)
+
+            results.append((receiver, result))
+
+        return results
+
+    def has_receivers_for(self, sender: t.Any) -> bool:
+        """Check if there is at least one receiver that will be called with the
+        given ``sender``. A receiver connected to :data:`ANY` will always be
+        called, regardless of sender. Does not check if weakly referenced
+        receivers are still live. See :meth:`receivers_for` for a stronger
+        search.
+
+        :param sender: Check for receivers connected to this sender, in addition
+            to those connected to :data:`ANY`.
+        """
+        if not self.receivers:
+            return False
+
+        if self._by_sender[ANY_ID]:
+            return True
+
+        if sender is ANY:
+            return False
+
+        return make_id(sender) in self._by_sender
+
+    def receivers_for(
+        self, sender: t.Any
+    ) -> c.Generator[c.Callable[..., t.Any], None, None]:
+        """Yield each receiver to be called for ``sender``, in addition to those
+        to be called for :data:`ANY`. Weakly referenced receivers that are not
+        live will be disconnected and skipped.
+
+        :param sender: Yield receivers connected to this sender, in addition
+            to those connected to :data:`ANY`.
+        """
+        # TODO: test receivers_for(ANY)
+        if not self.receivers:
+            return
+
+        sender_id = make_id(sender)
+
+        if sender_id in self._by_sender:
+            ids = self._by_sender[ANY_ID] | self._by_sender[sender_id]
+        else:
+            ids = self._by_sender[ANY_ID].copy()
+
+        for receiver_id in ids:
+            receiver = self.receivers.get(receiver_id)
+
+            if receiver is None:
+                continue
+
+            if isinstance(receiver, weakref.ref):
+                strong = receiver()
+
+                if strong is None:
+                    self._disconnect(receiver_id, ANY_ID)
+                    continue
+
+                yield strong
+            else:
+                yield receiver
+
+    def disconnect(self, receiver: c.Callable[..., t.Any], sender: t.Any = ANY) -> None:
+        """Disconnect ``receiver`` from being called when the signal is sent by
+        ``sender``.
+
+        :param receiver: A connected receiver callable.
+        :param sender: Disconnect from only this sender. By default, disconnect
+            from all senders.
+        """
+        sender_id: c.Hashable
+
+        if sender is ANY:
+            sender_id = ANY_ID
+        else:
+            sender_id = make_id(sender)
+
+        receiver_id = make_id(receiver)
+        self._disconnect(receiver_id, sender_id)
+
+        if (
+            "receiver_disconnected" in self.__dict__
+            and self.receiver_disconnected.receivers
+        ):
+            self.receiver_disconnected.send(self, receiver=receiver, sender=sender)
+
+    def _disconnect(self, receiver_id: c.Hashable, sender_id: c.Hashable) -> None:
+        if sender_id == ANY_ID:
+            if self._by_receiver.pop(receiver_id, None) is not None:
+                for bucket in self._by_sender.values():
+                    bucket.discard(receiver_id)
+
+            self.receivers.pop(receiver_id, None)
+        else:
+            self._by_sender[sender_id].discard(receiver_id)
+            self._by_receiver[receiver_id].discard(sender_id)
+
+    def _make_cleanup_receiver(
+        self, receiver_id: c.Hashable
+    ) -> c.Callable[[weakref.ref[c.Callable[..., t.Any]]], None]:
+        """Create a callback function to disconnect a weakly referenced
+        receiver when it is garbage collected.
+        """
+
+        def cleanup(ref: weakref.ref[c.Callable[..., t.Any]]) -> None:
+            # If the interpreter is shutting down, disconnecting can result in a
+            # weird ignored exception. Don't call it in that case.
+            if not sys.is_finalizing():
+                self._disconnect(receiver_id, ANY_ID)
+
+        return cleanup
+
+    def _make_cleanup_sender(
+        self, sender_id: c.Hashable
+    ) -> c.Callable[[weakref.ref[t.Any]], None]:
+        """Create a callback function to disconnect all receivers for a weakly
+        referenced sender when it is garbage collected.
+        """
+        assert sender_id != ANY_ID
+
+        def cleanup(ref: weakref.ref[t.Any]) -> None:
+            self._weak_senders.pop(sender_id, None)
+
+            for receiver_id in self._by_sender.pop(sender_id, ()):
+                self._by_receiver[receiver_id].discard(sender_id)
+
+        return cleanup
+
+    def _cleanup_bookkeeping(self) -> None:
+        """Prune unused sender/receiver bookkeeping. Not threadsafe.
+
+        Connecting & disconnecting leaves behind a small amount of bookkeeping
+        data. Typical workloads using Blinker, for example in most web apps,
+        Flask, CLI scripts, etc., are not adversely affected by this
+        bookkeeping.
+
+        With a long-running process performing dynamic signal routing with high
+        volume, e.g. connecting to function closures, senders are all unique
+        object instances. Doing all of this over and over may cause memory usage
+        to grow due to extraneous bookkeeping. (An empty ``set`` for each stale
+        sender/receiver pair.)
+
+        This method will prune that bookkeeping away, with the caveat that such
+        pruning is not threadsafe. The risk is that cleanup of a fully
+        disconnected receiver/sender pair occurs while another thread is
+        connecting that same pair. If you are in the highly dynamic, unique
+        receiver/sender situation that has lead you to this method, that failure
+        mode is perhaps not a big deal for you.
+        """
+        for mapping in (self._by_sender, self._by_receiver):
+            for ident, bucket in list(mapping.items()):
+                if not bucket:
+                    mapping.pop(ident, None)
+
+    def _clear_state(self) -> None:
+        """Disconnect all receivers and senders. Useful for tests."""
+        self._weak_senders.clear()
+        self.receivers.clear()
+        self._by_sender.clear()
+        self._by_receiver.clear()
+
+
+class NamedSignal(Signal):
+    """A named generic notification emitter. The name is not used by the signal
+    itself, but matches the key in the :class:`Namespace` that it belongs to.
+
+    :param name: The name of the signal within the namespace.
+    :param doc: The docstring for the signal.
+    """
+
+    def __init__(self, name: str, doc: str | None = None) -> None:
+        super().__init__(doc)
+
+        #: The name of this signal.
+        self.name: str = name
+
+    def __repr__(self) -> str:
+        base = super().__repr__()
+        return f"{base[:-1]}; {self.name!r}>"  # noqa: E702
+
+
+class Namespace(dict[str, NamedSignal]):
+    """A dict mapping names to signals."""
+
+    def signal(self, name: str, doc: str | None = None) -> NamedSignal:
+        """Return the :class:`NamedSignal` for the given ``name``, creating it
+        if required. Repeated calls with the same name return the same signal.
+
+        :param name: The name of the signal.
+        :param doc: The docstring of the signal.
+        """
+        if name not in self:
+            self[name] = NamedSignal(name, doc)
+
+        return self[name]
+
+
+class _PNamespaceSignal(t.Protocol):
+    def __call__(self, name: str, doc: str | None = None) -> NamedSignal: ...
+
+
+default_namespace: Namespace = Namespace()
+"""A default :class:`Namespace` for creating named signals. :func:`signal`
+creates a :class:`NamedSignal` in this namespace.
+"""
+
+signal: _PNamespaceSignal = default_namespace.signal
+"""Return a :class:`NamedSignal` in :data:`default_namespace` with the given
+``name``, creating it if required. Repeated calls with the same name return the
+same signal.
+"""

lib/python3.12/site-packages/einops-0.8.2.dist-info/INSTALLER

@@ -0,0 +1 @@
+pip

lib/python3.12/site-packages/einops-0.8.2.dist-info/METADATA

@@ -0,0 +1,394 @@
+Metadata-Version: 2.4
+Name: einops
+Version: 0.8.2
+Summary: A new flavour of deep learning operations
+Project-URL: Homepage, https://github.com/arogozhnikov/einops
+Author: Alex Rogozhnikov
+License: MIT
+License-File: LICENSE
+Keywords: deep learning,einops,machine learning,neural networks,scientific computations,tensor manipulation
+Classifier: Intended Audience :: Science/Research
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Programming Language :: Python :: 3
+Requires-Python: >=3.9
+Description-Content-Type: text/markdown
+
+
+<!--
+<a href='http://arogozhnikov.github.io/images/einops/einops_video.mp4' >
+<div align="center">
+  <img src="http://arogozhnikov.github.io/images/einops/einops_video.gif" alt="einops package examples" />
+  <br>
+  <small><a href='http://arogozhnikov.github.io/images/einops/einops_video.mp4'>This video in high quality (mp4)</a></small>
+  <br><br>
+</div>
+</a>
+-->
+
+<!-- this link magically rendered as video on github readme, unfortunately not in docs -->
+
+https://user-images.githubusercontent.com/6318811/177030658-66f0eb5d-e136-44d8-99c9-86ae298ead5b.mp4
+
+
+
+
+# einops 
+[![Run tests](https://github.com/arogozhnikov/einops/actions/workflows/run_tests.yml/badge.svg)](https://github.com/arogozhnikov/einops/actions/workflows/run_tests.yml)
+[![PyPI version](https://badge.fury.io/py/einops.svg)](https://badge.fury.io/py/einops)
+[![Documentation](https://img.shields.io/badge/documentation-link-blue.svg)](https://einops.rocks/)
+![Supported python versions](https://raw.githubusercontent.com/arogozhnikov/einops/main/docs/resources/python_badge.svg)
+
+
+Flexible and powerful tensor operations for readable and reliable code. <br />
+Supports numpy, pytorch, tensorflow, jax, and [others](#supported-frameworks).
+
+## Recent updates:
+
+- 0.8.0: tinygrad backend added, small fixes
+- 0.7.0: no-hassle `torch.compile`, support of [array api standard](https://data-apis.org/array-api/latest/API_specification/index.html) and more
+- 10'000🎉: github reports that more than 10k project use einops
+- einops 0.6.1: paddle backend added
+- einops 0.6 introduces [packing and unpacking](https://github.com/arogozhnikov/einops/blob/main/docs/4-pack-and-unpack.ipynb)
+- einops 0.5: einsum is now a part of einops
+- [Einops paper](https://openreview.net/pdf?id=oapKSVM2bcj) is accepted for oral presentation at ICLR 2022 (yes, it worth reading).
+  Talk recordings are [available](https://iclr.cc/virtual/2022/oral/6603)
+
+
+<details markdown="1">
+<summary>Previous updates</summary>
+- flax and oneflow backend added
+- torch.jit.script is supported for pytorch layers
+- powerful EinMix added to einops. [Einmix tutorial notebook](https://github.com/arogozhnikov/einops/blob/main/docs/3-einmix-layer.ipynb) 
+</details>
+
+<!--<div align="center">
+  <img src="http://arogozhnikov.github.io/images/einops/einops_logo_350x350.png" 
+  alt="einops package logo" width="250" height="250" />
+  <br><br>
+</div> -->
+
+
+## Tweets 
+
+> In case you need convincing arguments for setting aside time to learn about einsum and einops...
+[Tim Rocktäschel](https://twitter.com/_rockt/status/1230818967205425152)
+
+> Writing better code with PyTorch and einops 👌
+[Andrej Karpathy](https://twitter.com/karpathy/status/1290826075916779520)
+
+> Slowly but surely, einops is seeping in to every nook and cranny of my code. If you find yourself shuffling around bazillion dimensional tensors, this might change your life
+[Nasim Rahaman](https://twitter.com/nasim_rahaman/status/1216022614755463169)
+
+[More testimonials](https://einops.rocks/pages/testimonials/)
+
+
+## Contents
+
+- [Installation](#Installation)
+- [Documentation](https://einops.rocks/)
+- [Tutorial](#Tutorials)
+- [API micro-reference](#API)
+- [Why use einops](#Why-use-einops-notation)
+- [Supported frameworks](#Supported-frameworks)
+- [Citing](#Citing)
+- [Repository](https://github.com/arogozhnikov/einops) and [discussions](https://github.com/arogozhnikov/einops/discussions)
+
+## Installation  <a name="Installation"></a>
+
+Plain and simple:
+```bash
+pip install einops
+```
+
+(`uv pip install einops` works as well)
+
+## Tutorials <a name="Tutorials"></a>
+
+Tutorials are the most convenient way to see `einops` in action
+
+- part 1: [einops fundamentals](https://github.com/arogozhnikov/einops/blob/main/docs/1-einops-basics.ipynb)
+- part 2: [einops for deep learning](https://github.com/arogozhnikov/einops/blob/main/docs/2-einops-for-deep-learning.ipynb)
+- part 3: [packing and unpacking](https://github.com/arogozhnikov/einops/blob/main/docs/4-pack-and-unpack.ipynb)
+- part 4: [improve pytorch code with einops](http://einops.rocks/pytorch-examples.html)
+
+Kapil Sachdeva recorded a small [intro to einops](https://www.youtube.com/watch?v=xGy75Pjsqzo).
+
+## API <a name="API"></a>
+
+`einops` has a minimalistic yet powerful API.
+
+Three core operations provided ([einops tutorial](https://github.com/arogozhnikov/einops/blob/main/docs/)
+shows those cover stacking, reshape, transposition, squeeze/unsqueeze, repeat, tile, concatenate, view and numerous reductions)
+
+```python
+from einops import rearrange, reduce, repeat
+# rearrange elements according to the pattern
+output_tensor = rearrange(input_tensor, 't b c -> b c t')
+# combine rearrangement and reduction
+output_tensor = reduce(input_tensor, 'b c (h h2) (w w2) -> b h w c', 'mean', h2=2, w2=2)
+# copy along a new axis
+output_tensor = repeat(input_tensor, 'h w -> h w c', c=3)
+```
+
+Later additions to the family are `pack` and `unpack` functions (better than stack/split/concatenate):
+
+```python
+from einops import pack, unpack
+# pack and unpack allow reversibly 'packing' multiple tensors into one.
+# Packed tensors may be of different dimensionality:
+packed,  ps = pack([class_token_bc, image_tokens_bhwc, text_tokens_btc], 'b * c')
+class_emb_bc, image_emb_bhwc, text_emb_btc = unpack(transformer(packed), ps, 'b * c')
+```
+
+Finally, einops provides einsum with a support of multi-lettered names:
+
+```python
+from einops import einsum, pack, unpack
+# einsum is like ... einsum, generic and flexible dot-product
+# but 1) axes can be multi-lettered  2) pattern goes last 3) works with multiple frameworks
+C = einsum(A, B, 'b t1 head c, b t2 head c -> b head t1 t2')
+```
+
+### EinMix
+
+`EinMix` is a generic linear layer, perfect for MLP Mixers and similar architectures.
+
+### Layers
+
+Einops provides layers (`einops` keeps a separate version for each framework) that reflect corresponding functions
+
+```python
+from einops.layers.torch      import Rearrange, Reduce
+from einops.layers.tensorflow import Rearrange, Reduce
+from einops.layers.flax       import Rearrange, Reduce
+from einops.layers.paddle     import Rearrange, Reduce
+```
+
+<details markdown="1">
+<summary>Example of using layers within a pytorch model</summary>
+Example given for pytorch, but code in other frameworks is almost identical
+
+```python 
+from torch.nn import Sequential, Conv2d, MaxPool2d, Linear, ReLU
+from einops.layers.torch import Rearrange
+
+model = Sequential(
+    ...,
+    Conv2d(6, 16, kernel_size=5),
+    MaxPool2d(kernel_size=2),
+    # flattening without need to write forward
+    Rearrange('b c h w -> b (c h w)'),
+    Linear(16*5*5, 120),
+    ReLU(),
+    Linear(120, 10),
+)
+```
+
+No more flatten needed!
+
+Additionally, torch layers as those are script-able and compile-able.
+Operations [are torch.compile-able](https://github.com/arogozhnikov/einops/wiki/Using-torch.compile-with-einops),
+ but not script-able due to limitations of torch.jit.script.
+</details>
+
+
+
+
+## Naming <a name="Naming"></a>
+
+`einops` stands for Einstein-Inspired Notation for operations 
+(though "Einstein operations" is more attractive and easier to remember).
+
+Notation was loosely inspired by Einstein summation (in particular by `numpy.einsum` operation).
+
+## Why use `einops` notation?! <a name="Why-use-einops-notation"></a>
+
+
+### Semantic information (being verbose in expectations)
+
+```python
+y = x.view(x.shape[0], -1)
+y = rearrange(x, 'b c h w -> b (c h w)')
+```
+While these two lines are doing the same job in *some* context,
+the second one provides information about the input and output.
+In other words, `einops` focuses on interface: *what is the input and output*, not *how* the output is computed.
+
+The next operation looks similar:
+
+```python
+y = rearrange(x, 'time c h w -> time (c h w)')
+```
+but it gives the reader a hint:
+this is not an independent batch of images we are processing,
+but rather a sequence (video).
+
+Semantic information makes the code easier to read and maintain.
+
+### Convenient checks
+
+Reconsider the same example:
+
+```python
+y = x.view(x.shape[0], -1) # x: (batch, 256, 19, 19)
+y = rearrange(x, 'b c h w -> b (c h w)')
+```
+The second line checks that the input has four dimensions,
+but you can also specify particular dimensions.
+That's opposed to just writing comments about shapes since comments don't prevent mistakes,
+not tested, and without code review tend to be outdated
+```python
+y = x.view(x.shape[0], -1) # x: (batch, 256, 19, 19)
+y = rearrange(x, 'b c h w -> b (c h w)', c=256, h=19, w=19)
+```
+
+### Result is strictly determined
+
+Below we have at least two ways to define the depth-to-space operation
+```python
+# depth-to-space
+rearrange(x, 'b c (h h2) (w w2) -> b (c h2 w2) h w', h2=2, w2=2)
+rearrange(x, 'b c (h h2) (w w2) -> b (h2 w2 c) h w', h2=2, w2=2)
+```
+There are at least four more ways to do it. Which one is used by the framework?
+
+These details are ignored, since *usually* it makes no difference,
+but it can make a big difference (e.g. if you use grouped convolutions in the next stage),
+and you'd like to specify this in your code.
+
+
+### Uniformity
+
+```python
+reduce(x, 'b c (x dx) -> b c x', 'max', dx=2)
+reduce(x, 'b c (x dx) (y dy) -> b c x y', 'max', dx=2, dy=3)
+reduce(x, 'b c (x dx) (y dy) (z dz) -> b c x y z', 'max', dx=2, dy=3, dz=4)
+```
+These examples demonstrated that we don't use separate operations for 1d/2d/3d pooling,
+those are all defined in a uniform way.
+
+Space-to-depth and depth-to space are defined in many frameworks but how about width-to-height? Here you go:
+
+```python
+rearrange(x, 'b c h (w w2) -> b c (h w2) w', w2=2)
+```
+
+
+### Framework independent behavior
+
+Even simple functions are defined differently by different frameworks
+
+```python
+y = x.flatten() # or flatten(x)
+```
+
+Suppose `x`'s shape was `(3, 4, 5)`, then `y` has shape ...
+
+- numpy, pytorch, cupy, chainer, jax: `(60,)`
+- keras, tensorflow.layers, gluon: `(3, 20)`
+
+`einops` works the same way in all frameworks.
+
+
+### Independence of framework terminology
+
+Example: `tile` vs `repeat` causes lots of confusion. To copy image along width:
+```python
+np.tile(image, (1, 2))    # in numpy
+image.repeat(1, 2)        # pytorch's repeat ~ numpy's tile
+```
+
+With einops you don't need to decipher which axis was repeated:
+```python
+repeat(image, 'h w -> h (tile w)', tile=2)  # in numpy
+repeat(image, 'h w -> h (tile w)', tile=2)  # in pytorch
+repeat(image, 'h w -> h (tile w)', tile=2)  # in tf
+repeat(image, 'h w -> h (tile w)', tile=2)  # in jax
+repeat(image, 'h w -> h (tile w)', tile=2)  # in cupy
+... (etc.)
+```
+
+[Testimonials](https://einops.rocks/pages/testimonials/) provide users' perspective on the same question.
+
+
+## Supported frameworks <a name="Supported-frameworks"></a>
+
+Einops works with ...
+
+- [numpy](http://www.numpy.org/)
+- [pytorch](https://pytorch.org/)
+- [tensorflow](https://www.tensorflow.org/)
+- [jax](https://github.com/google/jax)
+- [cupy](https://github.com/cupy/cupy)
+- [flax](https://github.com/google/flax) (community)
+- [paddle](https://github.com/PaddlePaddle/Paddle) (community)
+- [oneflow](https://github.com/Oneflow-Inc/oneflow) (community)
+- [tinygrad](https://github.com/tinygrad/tinygrad) (community)
+- [pytensor](https://github.com/pymc-devs/pytensor) (community)
+
+
+
+```python
+from einops import rearrange  
+=> from einops.array_api import rearrange
+```
+
+But actually it is even better: einops can be used with *any* framework that supports
+[Python array API standard](https://data-apis.org/array-api/latest/API_specification/index.html),
+to name a few:
+
+- numpy >= 2.0
+- [MLX](https://github.com/ml-explore/mlx)  # yes, einops works with apple's framework
+- [pydata/sparse](https://github.com/pydata/sparse) >= 0.15 # and works with sparse tensors
+- [cubed](https://github.com/cubed-dev/cubed) # and with distributed tensors too
+- [quantco/ndonnx](https://github.com/Quantco/ndonnx)
+- jax
+- cupy
+- dask is supported via [array-api-compat](https://github.com/data-apis/array-api-compat)
+
+
+## Development
+
+Devcontainer is provided, this environment can be used locally, or on your server,
+or within github codespaces. 
+To start with devcontainers in vs code, clone repo, and click 'Reopen in Devcontainer'. 
+
+Starting from einops 0.8.1, einops distributes tests as a part of package.
+
+```bash
+# pip install einops pytest
+python -m einops.tests.run_tests numpy pytorch jax --pip-install
+```
+
+`numpy pytorch jax` is an _example_, any subset of testable frameworks can be provided.
+Every framework is tested against numpy, so it is a requirement for tests.
+
+Specifying `--pip-install` will install requirements in current virtualenv,
+and should be omitted if dependencies are installed locally.
+
+To build/test docs:
+
+```bash
+hatch run docs:serve  # Serving on http://localhost:8000/
+```
+
+
+## Citing einops <a name="Citing"></a>
+
+Please use the following bibtex record
+
+```text
+@inproceedings{
+    rogozhnikov2022einops,
+    title={Einops: Clear and Reliable Tensor Manipulations with Einstein-like Notation},
+    author={Alex Rogozhnikov},
+    booktitle={International Conference on Learning Representations},
+    year={2022},
+    url={https://openreview.net/forum?id=oapKSVM2bcj}
+}
+```
+
+
+## Supported python versions
+
+`einops` works with python 3.9 or later.

lib/python3.12/site-packages/einops-0.8.2.dist-info/RECORD

@@ -0,0 +1,58 @@
+einops-0.8.2.dist-info/INSTALLER,sha256=zuuue4knoyJ-UwPPXg8fezS7VCrXJQrAP7zeNuwvFQg,4
+einops-0.8.2.dist-info/METADATA,sha256=C9ZdsENoikT0opYIwculIcky9_928ji72GlEHhUBJlM,13724
+einops-0.8.2.dist-info/RECORD,,
+einops-0.8.2.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+einops-0.8.2.dist-info/licenses/LICENSE,sha256=MNmENkKW9R_67K1LAe4SfpUlDFBokY1LZvyWIGcj5DQ,1073
+einops/__init__.py,sha256=MwPcYWGHKsJE0Uxlhf7YMv_Sf6vu9W9Y98hy1mRM6zk,438
+einops/__pycache__/__init__.cpython-312.pyc,,
+einops/__pycache__/_backends.cpython-312.pyc,,
+einops/__pycache__/_torch_specific.cpython-312.pyc,,
+einops/__pycache__/array_api.cpython-312.pyc,,
+einops/__pycache__/einops.cpython-312.pyc,,
+einops/__pycache__/packing.cpython-312.pyc,,
+einops/__pycache__/parsing.cpython-312.pyc,,
+einops/_backends.py,sha256=v4snGd66fVWeK0M8e3lNJ9j7_qmd8RMRyJVUTQB5slc,22528
+einops/_torch_specific.py,sha256=QXgLFLDIgDa_Iqtv5gcucEeD3LHALYHJhaXRKgC5f4A,4329
+einops/array_api.py,sha256=7mQ-4yhlWtqmtvcTZyI-pmDOSDbk3qpSCExl4-Vm1-o,5236
+einops/einops.py,sha256=vmx5IWde_ePuzrPDLZZVw_-0Qsare1k8YRRaKt5ZL-8,38079
+einops/experimental/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+einops/experimental/__pycache__/__init__.cpython-312.pyc,,
+einops/experimental/__pycache__/indexing.cpython-312.pyc,,
+einops/experimental/indexing.py,sha256=yFFflW3-kV6_5PPJU7_jOJsJBCWCWlE4dGlu9gwSPXo,121
+einops/layers/__init__.py,sha256=jijwEY07Yr3rBx8N5zwZ-5qfbl-KX2BcRDK8NJFDaWA,3705
+einops/layers/__pycache__/__init__.cpython-312.pyc,,
+einops/layers/__pycache__/_einmix.cpython-312.pyc,,
+einops/layers/__pycache__/flax.cpython-312.pyc,,
+einops/layers/__pycache__/keras.cpython-312.pyc,,
+einops/layers/__pycache__/oneflow.cpython-312.pyc,,
+einops/layers/__pycache__/paddle.cpython-312.pyc,,
+einops/layers/__pycache__/tensorflow.cpython-312.pyc,,
+einops/layers/__pycache__/torch.cpython-312.pyc,,
+einops/layers/_einmix.py,sha256=MQr6rymfMw-nzFmhvaLCcee0o7Ai8a1lmluxndWDAjU,11075
+einops/layers/flax.py,sha256=Hk7L31wR2BHL63950-0hy3ivw4rc8NkgGW9jn1AKs-8,2518
+einops/layers/keras.py,sha256=jxD9ZgYLWsKMkCIwojbBNSrY6c1pfXYhE2AMZa0pNXA,217
+einops/layers/oneflow.py,sha256=djcKEZRJrWvxTuSsO4jJVZojieKiDPdUpP4SOpg_zTM,1864
+einops/layers/paddle.py,sha256=YCHffUB4ha7wG-xEOeR9XcTPl48Kr57k_nBEx-J6FVA,1907
+einops/layers/tensorflow.py,sha256=MjGApMbkb_gwAdGya7uGfPZ7olyETsRzbwymfJHii9E,3322
+einops/layers/torch.py,sha256=BV-PzjoijGU9RGea1HCFuIMx4yccicz767M0jJDIH6U,2405
+einops/packing.py,sha256=EZeGLgMRD78hB-0e80ubXvWHukh0_O_wJCf0Hbbdylo,7660
+einops/parsing.py,sha256=vNPYcAs54-KLX1RtbkypGpsSpujzVkj_dQkZlUinLYw,6805
+einops/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+einops/tests/__init__.py,sha256=CpSdrOGBSspTwuG6cAsYMGAwkTq0VuIBrehM7K-SYwU,3531
+einops/tests/__pycache__/__init__.cpython-312.pyc,,
+einops/tests/__pycache__/run_tests.cpython-312.pyc,,
+einops/tests/__pycache__/test_einsum.cpython-312.pyc,,
+einops/tests/__pycache__/test_examples.cpython-312.pyc,,
+einops/tests/__pycache__/test_layers.cpython-312.pyc,,
+einops/tests/__pycache__/test_ops.cpython-312.pyc,,
+einops/tests/__pycache__/test_other.cpython-312.pyc,,
+einops/tests/__pycache__/test_packing.cpython-312.pyc,,
+einops/tests/__pycache__/test_parsing.cpython-312.pyc,,
+einops/tests/run_tests.py,sha256=efJ2apHPOw1BKNIv5qZSNjHpgjSv0rJ4v9hbPbE7yNE,3072
+einops/tests/test_einsum.py,sha256=Nycwg4a-4_Vz7l5f7mj2KIQ7QVs8P2SQIy1rnDcMOG8,11022
+einops/tests/test_examples.py,sha256=6VVkHgov6hpEyoUGbMzFvRfoWTdP1Wr8RPji7G4pPlw,11510
+einops/tests/test_layers.py,sha256=m8RqmBojvVMgr_ET6mhcGpbfYPasg60jUYux7ZlXhSM,18667
+einops/tests/test_ops.py,sha256=4nxEGJWrmEG6cctYHGYuVVdEs5nUN9ziH7ukQYar44M,26884
+einops/tests/test_other.py,sha256=vkWSJr7sn-COewksivTPaLe-GAmzLt7IEZF9TMnt360,13521
+einops/tests/test_packing.py,sha256=AIBNiy8RftTj8dLv1AdNijkTXfrjCOTSw43fuD6Pfjc,10408
+einops/tests/test_parsing.py,sha256=-YAJeJlhDb3U6xYIDRrtm6p4MSNWogeH6l3-np1wUCA,4389

lib/python3.12/site-packages/einops-0.8.2.dist-info/WHEEL

@@ -0,0 +1,4 @@
+Wheel-Version: 1.0
+Generator: hatchling 1.27.0
+Root-Is-Purelib: true
+Tag: py3-none-any

lib/python3.12/site-packages/einops-0.8.2.dist-info/licenses/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2018 Alex Rogozhnikov
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

lib/python3.12/site-packages/functorch/__init__.py

@@ -0,0 +1,39 @@
+# Copyright (c) Facebook, Inc. and its affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the BSD-style license found in the
+# LICENSE file in the root directory of this source tree.
+import torch
+from torch._functorch.deprecated import (
+    combine_state_for_ensemble,
+    functionalize,
+    grad,
+    grad_and_value,
+    hessian,
+    jacfwd,
+    jacrev,
+    jvp,
+    make_functional,
+    make_functional_with_buffers,
+    vjp,
+    vmap,
+)
+
+# utilities. Maybe these should go in their own namespace in the future?
+from torch._functorch.make_functional import (
+    FunctionalModule,
+    FunctionalModuleWithBuffers,
+)
+
+# Was never documented
+from torch._functorch.python_key import make_fx
+
+
+# Top-level APIs. Please think carefully before adding something to the
+# top-level namespace:
+# - private helper functions should go into torch._functorch
+# - very experimental things should go into functorch.experimental
+# - compilation related things should go into functorch.compile
+
+
+__version__ = torch.__version__

lib/python3.12/site-packages/functorch/_src/aot_autograd/__init__.py

@@ -0,0 +1,8 @@
+# This file has moved to under torch/_functorch. It is not public API.
+# If you are not a PyTorch developer and you are relying on the following
+# imports, please file an issue.
+from torch._functorch.aot_autograd import (
+    aot_autograd_decompositions,
+    KNOWN_TYPES,
+    PytreeThunk,
+)

lib/python3.12/site-packages/functorch/_src/eager_transforms/__init__.py

@@ -0,0 +1,7 @@
+# This file has moved to under torch/_functorch. It is not public API.
+# If you are not a PyTorch developer and you are relying on the following
+# imports, please file an issue.
+from torch._functorch.eager_transforms import (
+    _assert_wrapped_functional,
+    _unwrap_functional_tensor,
+)

lib/python3.12/site-packages/functorch/_src/make_functional/__init__.py

@@ -0,0 +1,4 @@
+# This file has moved to under torch/_functorch. It is not public API.
+# If you are not a PyTorch developer and you are relying on the following
+# imports, please file an issue.
+from torch._functorch.make_functional import _swap_state

lib/python3.12/site-packages/functorch/_src/vmap/__init__.py

@@ -0,0 +1,16 @@
+# This file has moved to under torch/_functorch. It is not public API.
+# If you are not a PyTorch developer and you are relying on the following
+# imports, please file an issue.
+from torch._functorch.vmap import (
+    _add_batch_dim,
+    _broadcast_to_and_flatten,
+    _create_batched_inputs,
+    _get_name,
+    _process_batched_inputs,
+    _remove_batch_dim,
+    _unwrap_batched,
+    _validate_and_get_batch_size,
+    Tensor,
+    tree_flatten,
+    tree_unflatten,
+)

lib/python3.12/site-packages/functorch/compile/__init__.py

@@ -0,0 +1,30 @@
+from torch._functorch import config
+from torch._functorch.aot_autograd import (
+    aot_function,
+    aot_module,
+    aot_module_simplified,
+    compiled_function,
+    compiled_module,
+    get_aot_compilation_context,
+    get_aot_graph_name,
+    get_graph_being_compiled,
+    make_boxed_compiler,
+    make_boxed_func,
+)
+from torch._functorch.compilers import (
+    debug_compile,
+    default_decompositions,
+    draw_graph_compile,
+    memory_efficient_fusion,
+    nnc_jit,
+    nop,
+    print_compile,
+    ts_compile,
+)
+from torch._functorch.fx_minifier import minifier
+from torch._functorch.partitioners import (
+    default_partition,
+    draw_graph,
+    min_cut_rematerialization_partition,
+)
+from torch._functorch.python_key import pythonkey_decompose

lib/python3.12/site-packages/functorch/dim/__init__.py

@@ -0,0 +1,181 @@
+import dis
+import inspect
+from typing import Sequence, Union
+
+import functorch._C
+import torch
+from functorch._C import dim as _C
+
+from .tree_map import tree_flatten, tree_map
+from .wrap_type import wrap_type
+
+
+_C._patch_tensor_class()
+dims, DimList, dimlists = _C.dims, _C.DimList, _C.dimlists
+
+
+class DimensionMismatchError(Exception):
+    pass
+
+
+class DimensionBindError(Exception):
+    pass
+
+
+from . import op_properties
+
+
+# use dict to avoid writing C++ bindings for set
+pointwise = dict.fromkeys(op_properties.pointwise, True)
+
+use_c = True
+if not use_c:
+    from . import reference
+
+
+class _Tensor:
+    # fast path around slow wrapping/unwrapping logic for simply queries used
+    # by the implementation...
+
+    @property
+    def dims(self):
+        return tuple(d for d in self._levels if isinstance(d, Dim))
+
+    def dim(self):
+        return self.ndim
+
+    if use_c:
+        __torch_function__ = classmethod(_C.__torch_function__)
+        expand = _C._instancemethod(_C.expand)
+    else:
+        __torch_function__ = reference.__torch_function__
+        expand = reference.expand
+
+    index = _C._instancemethod(_C.index)
+
+    def __repr__(self):
+        tensor, levels, ndim = self._tensor, self._levels, self.ndim
+        return f"{tensor}\nwith dims={tuple(l + ndim if isinstance(l, int) else l for l in levels)} sizes={tuple(tensor.size())}"
+
+
+TensorLike = (_Tensor, torch.Tensor)
+
+
+class Dim(_C.Dim, _Tensor):
+    # note that _C.Dim comes before tensor because we want the Dim API for things like size to take precendence.
+    # Tensor defines format, but we want to print Dims with special formatting
+    __format__ = object.__format__
+
+
+class Tensor(_Tensor, _C.Tensor):
+    if not use_c:
+        from_batched = staticmethod(_C.Tensor_from_batched)
+    from_positional = staticmethod(_C.Tensor_from_positional)
+    sum = _C._instancemethod(_C.Tensor_sum)
+
+
+def cat(tensors, dim, new_dim):
+    n = dims()
+    return stack(tensors, n, dim).index([n, dim], new_dim)
+
+
+if use_c:
+    _wrap = _C._wrap
+
+    def _def(name, *args, **kwargs):
+        orig = getattr(torch.Tensor, name)
+        setattr(_Tensor, name, _C._instancemethod(_wrap(orig, *args, **kwargs)))
+
+    t__getitem__ = _C._instancemethod(_C.__getitem__)
+    stack = _C.stack
+    split = _C._instancemethod(_C.split)
+else:
+    _wrap, _def = reference._wrap, reference._def
+    t__getitem__ = reference.t__getitem__
+    stack = reference.stack
+    split = reference.split
+
+# note: there is no python reference
+t__setitem__ = _C._instancemethod(_C.__setitem__)
+# this is patched in the C API because otherwise torch.Tensor will
+# no longer be considered a sequence and things will break
+# torch.Tensor.__getitem__ = t__getitem__
+
+_Tensor.__getitem__ = t__getitem__
+# torch.Tensor.__setitem__ = t__setitem__
+_Tensor.__setitem__ = t__setitem__
+
+torch.Tensor.split = split
+_Tensor.split = split
+torch.Tensor.expand = _C._instancemethod(_C.expand)
+torch.Tensor.index = _C._instancemethod(_C.index)
+wrap_type(use_c, _Tensor, torch.Tensor, _Tensor.__torch_function__)
+del _Tensor.ndim
+
+if use_c:
+    _Tensor.order = _C._instancemethod(_C.order)
+else:
+    _Tensor.order = reference.positional
+
+_def("mean")
+_def("sum")
+_def("all")
+_def("amax")
+_def("amin")
+_def("aminmax")
+_def("any")
+_def("count_nonzero")
+_def("logsumexp")
+_def("nanmean")
+_def("nansum")
+_def("prod")
+_def("std", keepdim_offset=2)
+_def("var", keepdim_offset=2)
+_def("max", single_dim=True)
+_def("min", single_dim=True)
+_def("argmax", single_dim=True)
+_def("argmin", single_dim=True)
+_def("kthvalue", single_dim=True)
+_def("median", single_dim=True)
+_def("nanmedian", single_dim=True)
+_def("mode", single_dim=True)
+_def("sort", reduce=False)
+_def("argsort", reduce=False)
+_def("unbind", single_dim=True)
+_def("chunk", dim_offset=1, reduce=False)
+_def("cummax", single_dim=True, reduce=False)
+_def("cummin", single_dim=True, reduce=False)
+_def("cumprod", single_dim=True, reduce=False)
+_def("cumprod_", single_dim=True, reduce=False)
+_def("cumsum", single_dim=True, reduce=False)
+_def("cumsum_", single_dim=True, reduce=False)
+_def("logcumsumexp", single_dim=True, reduce=False)
+_def("renorm", dim_offset=1, single_dim=True, reduce=False)
+_def("softmax", single_dim=True, reduce=False)
+softmax = _wrap(torch.nn.functional.softmax, single_dim=True, reduce=False)
+
+# stuff to handle in the future, because they require special
+# binding logic for dims
+# cross
+# diag_embed
+# diagonal
+# diagonal_scatter
+# diff
+# nanquantile
+# quantile
+# roll
+# rot90
+# topk (new dimes on output)
+# should these all be subsumed by inplace indexing?
+# index_add_
+# index_add
+# index_copy
+# index_copy_
+# index_fill
+# index_fill_
+# index_select
+# scatter
+# scatter_
+# scatter_add
+# scatter_add_
+# scatter_reduce