optree 0.8.0

Optimized PyTree Utilities.

OpTree

Optimized PyTree Utilities.

---

Table of Contents

• Installation
• PyTrees
  • Tree Nodes and Leaves
    • Built-in PyTree Node Types
    • Registering a Container-like Custom Type as Non-leaf Nodes
    • Notes about the PyTree Type Registry
  • None is Non-leaf Node vs. None is Leaf
  • Key Ordering for Dictionaries
• Benchmark
  • Tree Flatten
  • Tree UnFlatten
  • Tree Flatten with Path
  • Tree Copy
  • Tree Map
  • Tree Map (nargs)
  • Tree Map with Path
  • Tree Map with Path (nargs)
• Changelog
• License

---

Installation

Install from PyPI ( / ):

pip3 install --upgrade optree

Install from conda-forge ():

conda install -c conda-forge optree

Install the latest version from GitHub:

pip3 install git+https://github.com/metaopt/optree.git#egg=optree

Or, clone this repo and install manually:

git clone --depth=1 https://github.com/metaopt/optree.git
cd optree
pip3 install .

Compiling from the source requires Python 3.7+, a compiler (gcc / clang / icc / cl.exe) supports C++20 and a cmake installation.

---

PyTrees

A PyTree is a recursive structure that can be an arbitrarily nested Python container (e.g., tuple, list, dict, OrderedDict, NamedTuple, etc.) or an opaque Python object. The key concepts of tree operations are tree flattening and its inverse (tree unflattening). Additional tree operations can be performed based on these two basic functions (e.g., tree_map = tree_unflatten ∘ map ∘ tree_flatten).

Tree flattening is traversing the entire tree in a left-to-right depth-first manner and returning the leaves of the tree in a deterministic order.

>>> tree = {'b': (2, [3, 4]), 'a': 1, 'c': 5, 'd': 6}
>>> optree.tree_flatten(tree)
([1, 2, 3, 4, 5, 6], PyTreeSpec({'a': *, 'b': (*, [*, *]), 'c': *, 'd': *}))
>>> optree.tree_flatten(1)
([1], PyTreeSpec(*))
>>> optree.tree_flatten(None)
([], PyTreeSpec(None))

This usually implies that the equal pytrees return equal lists of leaves and the same tree structure. See also section Key Ordering for Dictionaries.

>>> {'a': [1, 2], 'b': [3]} == {'b': [3], 'a': [1, 2]}
True
>>> optree.tree_leaves({'a': [1, 2], 'b': [3]}) == optree.tree_leaves({'b': [3], 'a': [1, 2]})
True
>>> optree.tree_structure({'a': [1, 2], 'b': [3]}) == optree.tree_structure({'b': [3], 'a': [1, 2]})
True

Tree Nodes and Leaves

A tree is a collection of non-leaf nodes and leaf nodes, where the leaf nodes have no children to flatten. optree.tree_flatten(...) will flatten the tree and return a list of leaf nodes while the non-leaf nodes will store in the tree specification.

Built-in PyTree Node Types

OpTree out-of-box supports the following Python container types in the registry:

• tuple
• list
• dict
• collections.namedtuple and its subclasses
• collections.OrderedDict
• collections.defaultdict
• collections.deque
• PyStructSequence types created by C API PyStructSequence_NewType

which are considered non-leaf nodes in the tree. Python objects that the type is not registered will be treated as leaf nodes. The registration lookup uses the is operator to determine whether the type is matched. So subclasses will need to explicitly register in the registry, otherwise, an object of that type will be considered as a leaf. The NoneType is a special case discussed in section None is non-leaf Node vs. None is Leaf.

Registering a Container-like Custom Type as Non-leaf Nodes

A container-like Python type can be registered in the type registry with a pair of functions that specify:

• flatten_func(container) -> (children, metadata, entries): convert an instance of the container type to a (children, metadata, entries) triple, where children is an iterable of subtrees and entries is an iterable of path entries of the container (e.g., indices or keys).
• unflatten_func(metadata, children) -> container: convert such a pair back to an instance of the container type.

The metadata is some necessary data apart from the children to reconstruct the container, e.g., the keys of the dictionary (the children are values).

The entries can be omitted (only returns a pair) or is optional to implement (returns None). If so, use range(len(children)) (i.e., flat indices) as path entries of the current node. The function signature can be flatten_func(container) -> (children, metadata) or flatten_func(container) -> (children, metadata, None).

The following examples show how to register custom types and utilize them for tree_flatten and tree_map. Please refer to section Notes about the PyTree Type Registry for more information.

# Registry a Python type with lambda functions
optree.register_pytree_node(
    set,
    # (set) -> (children, metadata, None)
    lambda s: (sorted(s), None, None),
    # (metadata, children) -> (set)
    lambda _, children: set(children),
    namespace='set',
)

# Register a Python type into a namespace
import torch

optree.register_pytree_node(
    torch.Tensor,
    # (tensor) -> (children, metadata)
    flatten_func=lambda tensor: (
        (tensor.cpu().numpy(),),
        dict(dtype=tensor.dtype, device=tensor.device, requires_grad=tensor.requires_grad),
    ),
    # (metadata, children) -> tensor
    unflatten_func=lambda metadata, children: torch.tensor(children[0], **metadata),
    namespace='torch2numpy',
)

>>> tree = {'weight': torch.ones(size=(1, 2)).cuda(), 'bias': torch.zeros(size=(2,))}
>>> tree
{'weight': tensor([[1., 1.]], device='cuda:0'), 'bias': tensor([0., 0.])}

# Flatten without specifying the namespace
>>> tree_flatten(tree)  # `torch.Tensor`s are leaf nodes
([tensor([0., 0.]), tensor([[1., 1.]], device='cuda:0')], PyTreeSpec({'bias': *, 'weight': *}))

# Flatten with the namespace
>>> leaves, treespec = optree.tree_flatten(tree, namespace='torch2numpy')
>>> leaves, treespec
(
    [array([0., 0.], dtype=float32), array([[1., 1.]], dtype=float32)],
    PyTreeSpec(
        {
            'bias': CustomTreeNode(Tensor[{'dtype': torch.float32, 'device': device(type='cpu'), 'requires_grad': False}], [*]),
            'weight': CustomTreeNode(Tensor[{'dtype': torch.float32, 'device': device(type='cuda', index=0), 'requires_grad': False}], [*])
        },
        namespace='torch2numpy'
    )
)

# `entries` are not defined and use `range(len(children))`
>>> optree.tree_paths(tree, namespace='torch2numpy')
[('bias', 0), ('weight', 0)]

# Unflatten back to a copy of the original object
>>> optree.tree_unflatten(treespec, leaves)
{'bias': tensor([0., 0.]), 'weight': tensor([[1., 1.]], device='cuda:0')}

Users can also extend the pytree registry by decorating the custom class and defining an instance method tree_flatten and a class method tree_unflatten.

from collections import UserDict

@optree.register_pytree_node_class(namespace='mydict')
class MyDict(UserDict):
    def tree_flatten(self):  # -> (children, metadata, entries)
        reversed_keys = sorted(self.keys(), reverse=True)
        return (
            [self[key] for key in reversed_keys],  # children
            reversed_keys,  # metadata
            reversed_keys,  # entries
        )

    @classmethod
    def tree_unflatten(cls, metadata, children):
        return cls(zip(metadata, children))

>>> tree = MyDict(b=4, a=(2, 3), c=MyDict({'d': 5, 'f': 6}))

# Flatten without specifying the namespace
>>> optree.tree_flatten_with_path(tree)  # `MyDict`s are leaf nodes
(
    [()],
    [MyDict(b=4, a=(2, 3), c=MyDict({'d': 5, 'f': 6}))],
    PyTreeSpec(*)
)

# Flatten with the namespace
>>> optree.tree_flatten_with_path(tree, namespace='mydict')
(
    [('c', 'f'), ('c', 'd'), ('b',), ('a', 0), ('a', 1)],
    [6, 5, 4, 2, 3],
    PyTreeSpec(
        CustomTreeNode(MyDict[['c', 'b', 'a']], [CustomTreeNode(MyDict[['f', 'd']], [*, *]), *, (*, *)]),
        namespace='mydict'
    )
)

Notes about the PyTree Type Registry

There are several key attributes of the pytree type registry:

1. The type registry is per-interpreter-dependent. This means registering a custom type in the registry affects all modules that use OpTree.

   - !!! WARNING !!!
     For safety reasons, a `namespace` must be specified while registering a custom type. It is
     used to isolate the behavior of flattening and unflattening a pytree node type. This is to
     prevent accidental collisions between different libraries that may register the same type.
   

2. The elements in the type registry are immutable. Users can neither register the same type twice in the same namespace (i.e., update the type registry), nor remove a type from the type registry. To update the behavior of an already registered type, simply register it again with another namespace.

3. Users cannot modify the behavior of already registered built-in types listed in Built-in PyTree Node Types, such as key order sorting for dict and collections.defaultdict.

4. Inherited subclasses are not implicitly registered. The registration lookup uses type(obj) is registered_type rather than isinstance(obj, registered_type). Users need to register the subclasses explicitly. To register all subclasses, it is easy to implement with metaclass or __init_subclass__, for example:

   from collections import UserDict
   
   @optree.register_pytree_node_class(namespace='mydict')
   class MyDict(UserDict):
       def __init_subclass__(cls):  # define this in the base class
           super().__init_subclass__()
           # Register a subclass to namespace 'mydict'
           optree.register_pytree_node_class(cls, namespace='mydict')
   
       def tree_flatten(self):  # -> (children, metadata, entries)
           reversed_keys = sorted(self.keys(), reverse=True)
           return (
               [self[key] for key in reversed_keys],  # children
               reversed_keys,  # metadata
               reversed_keys,  # entries
           )
   
       @classmethod
       def tree_unflatten(cls, metadata, children):
           return cls(zip(metadata, children))
   
   # Subclasses will be automatically registered in namespace 'mydict'
   class MyAnotherDict(MyDict):
       pass
   

   >>> tree = MyDict(b=4, a=(2, 3), c=MyAnotherDict({'d': 5, 'f': 6}))
   >>> optree.tree_flatten_with_path(tree, namespace='mydict')
   (
       [('c', 'f'), ('c', 'd'), ('b',), ('a', 0), ('a', 1)],
       [6, 5, 4, 2, 3],
       PyTreeSpec(
           CustomTreeNode(MyDict[['c', 'b', 'a']], [CustomTreeNode(MyAnotherDict[['f', 'd']], [*, *]), *, (*, *)]),
           namespace='mydict'
       )
   )
   

5. Be careful about the potential infinite recursion of the custom flatten function. The returned children from the custom flatten function are considered subtrees. They will be further flattened recursively. The children can have the same type as the current node. Users must design their termination condition carefully.

   import numpy as np
   import torch
   
   optree.register_pytree_node(
       np.ndarray,
       # Children are nest lists of Python objects
       lambda array: (np.atleast_1d(array).tolist(), array.ndim == 0),
       lambda scalar, rows: np.asarray(rows) if not scalar else np.asarray(rows[0]),
       namespace='numpy1',
   )
   
   optree.register_pytree_node(
       np.ndarray,
       # Children are Python objects
       lambda array: (
           list(array.ravel()),  # list(1DArray[T]) -> List[T]
           dict(shape=array.shape, dtype=array.dtype)
       ),
       lambda metadata, children: np.asarray(children, dtype=metadata['dtype']).reshape(metadata['shape']),
       namespace='numpy2',
   )
   
   optree.register_pytree_node(
       np.ndarray,
       # Returns a list of `np.ndarray`s without termination condition
       lambda array: ([array.ravel()], array.dtype),
       lambda shape, children: children[0].reshape(shape),
       namespace='numpy3',
   )
   
   optree.register_pytree_node(
       torch.Tensor,
       # Children are nest lists of Python objects
       lambda tensor: (torch.atleast_1d(tensor).tolist(), tensor.ndim == 0),
       lambda scalar, rows: torch.tensor(rows) if not scalar else torch.tensor(rows[0])),
       namespace='torch1',
   )
   
   optree.register_pytree_node(
       torch.Tensor,
       # Returns a list of `torch.Tensor`s without termination condition
       lambda tensor: (
           list(tensor.view(-1)),  # list(1DTensor[T]) -> List[0DTensor[T]] (STILL TENSORS!)
           tensor.shape
       ),
       lambda shape, children: torch.stack(children).reshape(shape),
       namespace='torch2',
   )
   

   >>> optree.tree_flatten(np.arange(9).reshape(3, 3), namespace='numpy1')
   (
       [0, 1, 2, 3, 4, 5, 6, 7, 8],
       PyTreeSpec(
           CustomTreeNode(ndarray[False], [[*, *, *], [*, *, *], [*, *, *]]),
           namespace='numpy1'
       )
   )
   # Implicitly casts `float`s to `np.float64`
   >>> optree.tree_map(lambda x: x + 1.5, np.arange(9).reshape(3, 3), namespace='numpy1')
   array([[1.5, 2.5, 3.5],
          [4.5, 5.5, 6.5],
          [7.5, 8.5, 9.5]])
   
   >>> optree.tree_flatten(np.arange(9).reshape(3, 3), namespace='numpy2')
   (
       [0, 1, 2, 3, 4, 5, 6, 7, 8],
       PyTreeSpec(
           CustomTreeNode(ndarray[{'shape': (3, 3), 'dtype': dtype('int64')}], [*, *, *, *, *, *, *, *, *]),
           namespace='numpy2'
       )
   )
   # Explicitly casts `float`s to `np.int64`
   >>> optree.tree_map(lambda x: x + 1.5, np.arange(9).reshape(3, 3), namespace='numpy2')
   array([[1, 2, 3],
          [4, 5, 6],
          [7, 8, 9]])
   
   # Children are also `np.ndarray`s, recurse without termination condition.
   >>> optree.tree_flatten(np.arange(9).reshape(3, 3), namespace='numpy3')
   Traceback (most recent call last):
       ...
   RecursionError: Maximum recursion depth exceeded during flattening the tree.
   
   >>> optree.tree_flatten(torch.arange(9).reshape(3, 3), namespace='torch1')
   (
       [0, 1, 2, 3, 4, 5, 6, 7, 8],
       PyTreeSpec(
           CustomTreeNode(Tensor[False], [[*, *, *], [*, *, *], [*, *, *]]),
           namespace='torch1'
       )
   )
   # Implicitly casts `float`s to `torch.float32`
   >>> optree.tree_map(lambda x: x + 1.5, torch.arange(9).reshape(3, 3), namespace='torch1')
   tensor([[1.5000, 2.5000, 3.5000],
           [4.5000, 5.5000, 6.5000],
           [7.5000, 8.5000, 9.5000]])
   
   # Children are also `torch.Tensor`s, recurse without termination condition.
   >>> optree.tree_flatten(torch.arange(9).reshape(3, 3), namespace='torch2')
   Traceback (most recent call last):
       ...
   RecursionError: Maximum recursion depth exceeded during flattening the tree.
   

None is Non-leaf Node vs. None is Leaf

The None object is a special object in the Python language. It serves some of the same purposes as null (a pointer does not point to anything) in other programming languages, which denotes a variable is empty or marks default parameters. However, the None object is a singleton object rather than a pointer. It may also serve as a sentinel value. In addition, if a function has returned without any return value or the return statement is omitted, the function will also implicitly return the None object.

By default, the None object is considered a non-leaf node in the tree with arity 0, i.e., a non-leaf node that has no children. This is like the behavior of an empty tuple. While flattening a tree, it will remain in the tree structure definitions rather than in the leaves list.

>>> tree = {'b': (2, [3, 4]), 'a': 1, 'c': None, 'd': 5}
>>> optree.tree_flatten(tree)
([1, 2, 3, 4, 5], PyTreeSpec({'a': *, 'b': (*, [*, *]), 'c': None, 'd': *}))
>>> optree.tree_flatten(tree, none_is_leaf=True)
([1, 2, 3, 4, None, 5], PyTreeSpec({'a': *, 'b': (*, [*, *]), 'c': *, 'd': *}, NoneIsLeaf))
>>> optree.tree_flatten(1)
([1], PyTreeSpec(*))
>>> optree.tree_flatten(None)
([], PyTreeSpec(None))
>>> optree.tree_flatten(None, none_is_leaf=True)
([None], PyTreeSpec(*, NoneIsLeaf))

OpTree provides a keyword argument none_is_leaf to determine whether to consider the None object as a leaf, like other opaque objects. If none_is_leaf=True, the None object will place in the leaves list. Otherwise, the None object will remain in the tree specification (structure).

>>> import torch

>>> linear = torch.nn.Linear(in_features=3, out_features=2, bias=False)
>>> linear._parameters  # a container has None
OrderedDict([
    ('weight', Parameter containing:
               tensor([[-0.6677,  0.5209,  0.3295],
                       [-0.4876, -0.3142,  0.1785]], requires_grad=True)),
    ('bias', None)
])

>>> optree.tree_map(torch.zeros_like, linear._parameters)
OrderedDict([
    ('weight', tensor([[0., 0., 0.],
                       [0., 0., 0.]])),
    ('bias', None)
])

>>> optree.tree_map(torch.zeros_like, linear._parameters, none_is_leaf=True)
Traceback (most recent call last):
    ...
TypeError: zeros_like(): argument 'input' (position 1) must be Tensor, not NoneType

>>> optree.tree_map(lambda t: torch.zeros_like(t) if t is not None else 0, linear._parameters, none_is_leaf=True)
OrderedDict([
    ('weight', tensor([[0., 0., 0.],
                       [0., 0., 0.]])),
    ('bias', 0)
])

Key Ordering for Dictionaries

The built-in Python dictionary (i.e., builtins.dict) is an unordered mapping that holds the keys and values. The leaves of a dictionary are the values. Although since Python 3.6, the built-in dictionary is insertion ordered (PEP 468). The dictionary equality operator (==) does not check for key ordering. To ensure that "equal dict" implies "equal ordering of leaves", the order of values of the dictionary is sorted by the keys. This behavior is also applied to collections.defaultdict.

>>> optree.tree_flatten({'a': [1, 2], 'b': [3]})
([1, 2, 3], PyTreeSpec({'a': [*, *], 'b': [*]}))
>>> optree.tree_flatten({'b': [3], 'a': [1, 2]})
([1, 2, 3], PyTreeSpec({'a': [*, *], 'b': [*]}))

Note that there are no restrictions on the dict to require the keys are comparable (sortable). There can be multiple types of keys in the dictionary. The keys are sorted in ascending order by key=lambda k: k first if capable otherwise fallback to key=lambda k: (k.__class__.__qualname__, k). This handles most cases.

>>> sorted({1: 2, 1.5: 1}.keys())
[1, 1.5]
>>> sorted({'a': 3, 1: 2, 1.5: 1}.keys())
Traceback (most recent call last):
    ...
TypeError: '<' not supported between instances of 'int' and 'str'
>>> sorted({'a': 3, 1: 2, 1.5: 1}.keys(), key=lambda k: (k.__class__.__qualname__, k))
[1.5, 1, 'a']

If users want to keep the values in the insertion order, they should use collections.OrderedDict, which will take the order of keys under consideration:

>>> OrderedDict([('a', [1, 2]), ('b', [3])]) == OrderedDict([('b', [3]), ('a', [1, 2])])
False
>>> optree.tree_flatten(OrderedDict([('a', [1, 2]), ('b', [3])]))
([1, 2, 3], PyTreeSpec(OrderedDict([('a', [*, *]), ('b', [*])])))
>>> optree.tree_flatten(OrderedDict([('b', [3]), ('a', [1, 2])]))
([3, 1, 2], PyTreeSpec(OrderedDict([('b', [*]), ('a', [*, *])])))

---

Benchmark

We benchmark the performance of:

• tree flatten
• tree unflatten
• tree copy (i.e., unflatten(flatten(...)))
• tree map

compared with the following libraries:

• OpTree (@v0.8.0)
• JAX XLA (jax[cpu] == 0.4.6)
• PyTorch (torch == 1.13.1)
• DM-Tree (dm-tree == 0.1.8)

All results are reported on a workstation with an AMD Ryzen 9 5950X CPU @ 4.45GHz in an isolated virtual environment with Python 3.10.9. Run with the following commands:

conda create --name optree-benchmark anaconda::python=3.10 --yes --no-default-packages
conda activate optree-benchmark
python3 -m pip install --editable '.[benchmark]' --extra-index-url https://download.pytorch.org/whl/cpu
python3 benchmark.py --number=10000 --repeat=5

The test inputs are nested containers (i.e., pytrees) extracted from torch.nn.Module objects. They are:

tiny_mlp = nn.Sequential(
    nn.Linear(1, 1, bias=True),
    nn.BatchNorm1d(1, affine=True, track_running_stats=True),
    nn.ReLU(),
    nn.Linear(1, 1, bias=False),
    nn.Sigmoid(),
)

and AlexNet, ResNet18, ResNet34, ResNet50, ResNet101, ResNet152, VisionTransformerH14 (ViT-H/14), and SwinTransformerB (Swin-B) from torchvsion. Please refer to benchmark.py for more details.

Tree Flatten

Module	Nodes	OpTree (μs)	JAX XLA (μs)	PyTorch (μs)	DM-Tree (μs)	Speedup (J / O)	Speedup (P / O)	Speedup (D / O)
TinyMLP	53	30.18	68.69	577.79	31.55	2.28	19.15	1.05
AlexNet	188	97.68	242.74	2102.67	118.22	2.49	21.53	1.21
ResNet18	698	346.24	787.04	7769.05	407.51	2.27	22.44	1.18
ResNet34	1242	663.70	1431.62	13989.08	712.72	2.16	21.08	1.07
ResNet50	1702	882.40	1906.07	19243.43	966.05	2.16	21.81	1.09
ResNet101	3317	1847.35	3953.69	39870.71	2031.28	2.14	21.58	1.10
ResNet152	4932	2678.84	5588.23	56023.10	2874.76	2.09	20.91	1.07
ViT-H/14	3420	1947.77	4467.48	40057.71	2195.20	2.29	20.57	1.13
Swin-B	2881	1763.83	3985.11	35818.71	1968.06	2.26	20.31	1.12
					Average	2.24	21.04	1.11

Tree UnFlatten

Module	Nodes	OpTree (μs)	JAX XLA (μs)	PyTorch (μs)	DM-Tree (μs)	Speedup (J / O)	Speedup (P / O)	Speedup (D / O)
TinyMLP	53	54.91	134.31	234.37	913.43	2.45	4.27	16.64
AlexNet	188	210.76	565.19	929.61	3808.84	2.68	4.41	18.07
ResNet18	698	703.22	1727.14	3184.54	11643.08	2.46	4.53	16.56
ResNet34	1242	1312.01	3147.73	5762.68	20852.70	2.40	4.39	15.89
ResNet50	1702	1758.62	4177.30	7891.72	27874.16	2.38	4.49	15.85
ResNet101	3317	3753.81	8226.49	15362.37	53974.51	2.19	4.09	14.38
ResNet152	4932	5313.30	12205.85	24068.88	80256.68	2.30	4.53	15.10
ViT-H/14	3420	3994.53	10016.00	17411.04	66000.54	2.51	4.36	16.52
Swin-B	2881	3584.82	8940.27	15582.13	56003.34	2.49	4.35	15.62
					Average	2.42	4.38	16.07

Tree Flatten with Path

Module	Nodes	OpTree (μs)	JAX XLA (μs)	PyTorch (μs)	DM-Tree (μs)	Speedup (J / O)	Speedup (P / O)	Speedup (D / O)
TinyMLP	53	35.55	522.00	N/A	915.30	14.68	N/A	25.75
AlexNet	188	113.89	2005.85	N/A	3503.25	17.61	N/A	30.76
ResNet18	698	432.31	7052.73	N/A	12239.45	16.31	N/A	28.31
ResNet34	1242	812.18	12657.12	N/A	21703.50	15.58	N/A	26.72
ResNet50	1702	1105.15	17173.43	N/A	29293.02	15.54	N/A	26.51
ResNet101	3317	2182.68	33455.81	N/A	56810.61	15.33	N/A	26.03
ResNet152	4932	3272.97	49550.72	N/A	84535.23	15.14	N/A	25.83
ViT-H/14	3420	2287.13	37485.28	N/A	63024.61	16.39	N/A	27.56
Swin-B	2881	2092.16	33942.08	N/A	52744.88	16.22	N/A	25.21
					Average	15.87	N/A	26.96

Tree Copy

Module	Nodes	OpTree (μs)	JAX XLA (μs)	PyTorch (μs)	DM-Tree (μs)	Speedup (J / O)	Speedup (P / O)	Speedup (D / O)
TinyMLP	53	91.28	211.31	833.84	952.72	2.31	9.13	10.44
AlexNet	188	320.20	825.66	3118.08	3938.46	2.58	9.74	12.30
ResNet18	698	1113.83	2578.31	11325.44	12068.00	2.31	10.17	10.83
ResNet34	1242	2050.00	4836.56	20324.52	22749.73	2.36	9.91	11.10
ResNet50	1702	2897.93	6121.16	27563.39	28840.04	2.11	9.51	9.95
ResNet101	3317	5456.58	12306.26	53733.62	56140.12	2.26	9.85	10.29
ResNet152	4932	8044.33	18873.23	79896.03	83215.06	2.35	9.93	10.34
ViT-H/14	3420	6046.78	14451.37	58204.01	70966.61	2.39	9.63	11.74
Swin-B	2881	5173.48	13174.36	51701.17	60053.21	2.55	9.99	11.61
					Average	2.36	9.76	10.96

Tree Map

Module	Nodes	OpTree (μs)	JAX XLA (μs)	PyTorch (μs)	DM-Tree (μs)	Speedup (J / O)	Speedup (P / O)	Speedup (D / O)
TinyMLP	53	99.25	229.26	848.37	968.28	2.31	8.55	9.76
AlexNet	188	332.74	853.33	3142.39	3992.22	2.56	9.44	12.00
ResNet18	698	1190.94	2760.28	11399.95	12294.02	2.32	9.57	10.32
ResNet34	1242	2286.53	4925.70	20423.57	23204.74	2.15	8.93	10.15
ResNet50	1702	2968.51	6622.94	27807.01	29259.40	2.23	9.37	9.86
ResNet101	3317	5851.06	13132.59	53999.13	57251.12	2.24	9.23	9.78
ResNet152	4932	8682.55	19346.59	80462.95	84364.39	2.23	9.27	9.72
ViT-H/14	3420	6695.68	16045.45	58313.07	68415.82	2.40	8.71	10.22
Swin-B	2881	5747.50	13757.05	52229.81	61017.78	2.39	9.09	10.62
					Average	2.32	9.13	10.27

Tree Map (nargs)

Module	Nodes	OpTree (μs)	JAX XLA (μs)	PyTorch (μs)	DM-Tree (μs)	Speedup (J / O)	Speedup (P / O)	Speedup (D / O)
TinyMLP	53	133.87	344.99	N/A	3599.07	2.58	N/A	26.89
AlexNet	188	445.41	1310.77	N/A	14207.10	2.94	N/A	31.90
ResNet18	698	1599.16	4239.56	N/A	49255.49	2.65	N/A	30.80
ResNet34	1242	3066.14	8115.79	N/A	88568.31	2.65	N/A	28.89
ResNet50	1702	3951.48	10557.52	N/A	127232.92	2.67	N/A	32.20
ResNet101	3317	7801.80	20208.53	N/A	235961.43	2.59	N/A	30.24
ResNet152	4932	11489.21	29375.98	N/A	349007.54	2.56	N/A	30.38
ViT-H/14	3420	8319.66	23204.11	N/A	266190.21	2.79	N/A	32.00
Swin-B	2881	7259.47	20098.17	N/A	226166.17	2.77	N/A	31.15
					Average	2.69	N/A	30.49

Tree Map with Path

Module	Nodes	OpTree (μs)	JAX XLA (μs)	PyTorch (μs)	DM-Tree (μs)	Speedup (J / O)	Speedup (P / O)	Speedup (D / O)
TinyMLP	53	104.70	703.83	N/A	1998.00	6.72	N/A	19.08
AlexNet	188	352.30	2668.73	N/A	7681.19	7.58	N/A	21.80
ResNet18	698	1289.51	9342.79	N/A	25497.31	7.25	N/A	19.77
ResNet34	1242	2366.46	16746.52	N/A	45254.59	7.08	N/A	19.12
ResNet50	1702	3399.11	23574.46	N/A	60494.27	6.94	N/A	17.80
ResNet101	3317	6329.82	43955.95	N/A	118725.60	6.94	N/A	18.76
ResNet152	4932	9307.87	64777.45	N/A	174764.97	6.96	N/A	18.78
ViT-H/14	3420	6705.10	48862.92	N/A	139617.21	7.29	N/A	20.82
Swin-B	2881	5780.20	41703.04	N/A	115003.61	7.21	N/A	19.90
					Average	7.11	N/A	19.54

Tree Map with Path (nargs)

Module	Nodes	OpTree (μs)	JAX XLA (μs)	PyTorch (μs)	DM-Tree (μs)	Speedup (J / O)	Speedup (P / O)	Speedup (D / O)
TinyMLP	53	138.19	828.53	N/A	3599.32	6.00	N/A	26.05
AlexNet	188	461.91	3138.59	N/A	14069.17	6.79	N/A	30.46
ResNet18	698	1702.79	10890.25	N/A	49456.32	6.40	N/A	29.04
ResNet34	1242	3115.89	19356.46	N/A	88955.96	6.21	N/A	28.55
ResNet50	1702	4422.25	26205.69	N/A	121569.30	5.93	N/A	27.49
ResNet101	3317	8334.83	50909.37	N/A	241862.38	6.11	N/A	29.02
ResNet152	4932	12208.52	75327.94	N/A	351472.89	6.17	N/A	28.79
ViT-H/14	3420	9320.75	56869.65	N/A	266430.78	6.10	N/A	28.58
Swin-B	2881	7472.11	49260.03	N/A	233154.60	6.59	N/A	31.20
					Average	6.26	N/A	28.80

---

Changelog

See CHANGELOG.md.

---

License

OpTree is released under the Apache License 2.0.

OpTree is heavily based on JAX's implementation of the PyTree utility, with deep refactoring and several improvements. The original licenses can be found at JAX's Apache License 2.0 and Tensorflow's Apache License 2.0.