• Stars
    star
    1,388
  • Rank 33,862 (Top 0.7 %)
  • Language
    Jupyter Notebook
  • License
    BSD 3-Clause "New...
  • Created over 3 years ago
  • Updated 3 months ago

Reviews

There are no reviews yet. Be the first to send feedback to the community and the maintainers!

Repository Details

functorch is JAX-like composable function transforms for PyTorch.

functorch

Why functorch? | Install guide | Transformations | Documentation | Future Plans

This library is currently under heavy development - if you have suggestions on the API or use-cases you'd like to be covered, please open an github issue or reach out. We'd love to hear about how you're using the library.

functorch is JAX-like composable function transforms for PyTorch.

It aims to provide composable vmap and grad transforms that work with PyTorch modules and PyTorch autograd with good eager-mode performance.

In addition, there is experimental functionality to trace through these transformations using FX in order to capture the results of these transforms ahead of time. This would allow us to compile the results of vmap or grad to improve performance.

Why composable function transforms?

There are a number of use cases that are tricky to do in PyTorch today:

  • computing per-sample-gradients (or other per-sample quantities)
  • running ensembles of models on a single machine
  • efficiently batching together tasks in the inner-loop of MAML
  • efficiently computing Jacobians and Hessians
  • efficiently computing batched Jacobians and Hessians

Composing vmap, grad, vjp, and jvp transforms allows us to express the above without designing a separate subsystem for each. This idea of composable function transforms comes from the JAX framework.

Install

There are two ways to install functorch:

  1. functorch from source
  2. functorch beta (compatible with recent PyTorch releases)

We recommend trying out the functorch beta first.

Installing functorch from source

Click to expand

Using Colab

Follow the instructions in this Colab notebook

Locally

As of 9/21/2022, functorch comes installed alongside a nightly PyTorch binary. Please install a Preview (nightly) PyTorch binary; see https://pytorch.org/ for instructions.

Once you've done that, run a quick sanity check in Python:

import torch
from functorch import vmap
x = torch.randn(3)
y = vmap(torch.sin)(x)
assert torch.allclose(y, x.sin())

functorch development setup

As of 9/21/2022, functorch comes installed alongside PyTorch and is in the PyTorch source tree. Please install PyTorch from source, then, you will be able to import functorch.

Try to run some tests to make sure all is OK:

pytest test/test_vmap.py -v
pytest test/test_eager_transforms.py -v

AOTAutograd has some additional optional requirements. You can install them via:

pip install networkx

To run functorch tests, please install our test dependencies (expecttest, pyyaml).

Installing functorch beta (compatible with recent PyTorch releases)

Click to expand

Using Colab

Follow the instructions here

pip

Prerequisite: Install PyTorch

pip install functorch

Finally, run a quick sanity check in python:

import torch
from functorch import vmap
x = torch.randn(3)
y = vmap(torch.sin)(x)
assert torch.allclose(y, x.sin())

What are the transforms?

Right now, we support the following transforms:

  • grad, vjp, jvp,
  • jacrev, jacfwd, hessian
  • vmap

Furthermore, we have some utilities for working with PyTorch modules.

  • make_functional(model)
  • make_functional_with_buffers(model)

vmap

Note: vmap imposes restrictions on the code that it can be used on. For more details, please read its docstring.

vmap(func)(*inputs) is a transform that adds a dimension to all Tensor operations in func. vmap(func) returns a new function that maps func over some dimension (default: 0) of each Tensor in inputs.

vmap is useful for hiding batch dimensions: one can write a function func that runs on examples and then lift it to a function that can take batches of examples with vmap(func), leading to a simpler modeling experience:

from functorch import vmap
batch_size, feature_size = 3, 5
weights = torch.randn(feature_size, requires_grad=True)

def model(feature_vec):
    # Very simple linear model with activation
    assert feature_vec.dim() == 1
    return feature_vec.dot(weights).relu()

examples = torch.randn(batch_size, feature_size)
result = vmap(model)(examples)

grad

grad(func)(*inputs) assumes func returns a single-element Tensor. It compute the gradients of the output of func w.r.t. to inputs[0].

from functorch import grad
x = torch.randn([])
cos_x = grad(lambda x: torch.sin(x))(x)
assert torch.allclose(cos_x, x.cos())

# Second-order gradients
neg_sin_x = grad(grad(lambda x: torch.sin(x)))(x)
assert torch.allclose(neg_sin_x, -x.sin())

When composed with vmap, grad can be used to compute per-sample-gradients:

from functorch import vmap
batch_size, feature_size = 3, 5

def model(weights,feature_vec):
    # Very simple linear model with activation
    assert feature_vec.dim() == 1
    return feature_vec.dot(weights).relu()

def compute_loss(weights, example, target):
    y = model(weights, example)
    return ((y - target) ** 2).mean()  # MSELoss

weights = torch.randn(feature_size, requires_grad=True)
examples = torch.randn(batch_size, feature_size)
targets = torch.randn(batch_size)
inputs = (weights,examples, targets)
grad_weight_per_example = vmap(grad(compute_loss), in_dims=(None, 0, 0))(*inputs)

vjp

The vjp transform applies func to inputs and returns a new function that computes vjps given some cotangents Tensors.

from functorch import vjp
outputs, vjp_fn = vjp(func, inputs); vjps = vjp_fn(*cotangents)

jvp

The jvp transforms computes Jacobian-vector-products and is also known as "forward-mode AD". It is not a higher-order function unlike most other transforms, but it returns the outputs of func(inputs) as well as the jvps.

from functorch import jvp
x = torch.randn(5)
y = torch.randn(5)
f = lambda x, y: (x * y)
_, output = jvp(f, (x, y), (torch.ones(5), torch.ones(5)))
assert torch.allclose(output, x + y)

jacrev, jacfwd, and hessian

The jacrev transform returns a new function that takes in x and returns the Jacobian of torch.sin with respect to x using reverse-mode AD.

from functorch import jacrev
x = torch.randn(5)
jacobian = jacrev(torch.sin)(x)
expected = torch.diag(torch.cos(x))
assert torch.allclose(jacobian, expected)

Use jacrev to compute the jacobian. This can be composed with vmap to produce batched jacobians:

x = torch.randn(64, 5)
jacobian = vmap(jacrev(torch.sin))(x)
assert jacobian.shape == (64, 5, 5)

jacfwd is a drop-in replacement for jacrev that computes Jacobians using forward-mode AD:

from functorch import jacfwd
x = torch.randn(5)
jacobian = jacfwd(torch.sin)(x)
expected = torch.diag(torch.cos(x))
assert torch.allclose(jacobian, expected)

Composing jacrev with itself or jacfwd can produce hessians:

def f(x):
  return x.sin().sum()

x = torch.randn(5)
hessian0 = jacrev(jacrev(f))(x)
hessian1 = jacfwd(jacrev(f))(x)

The hessian is a convenience function that combines jacfwd and jacrev:

from functorch import hessian

def f(x):
  return x.sin().sum()

x = torch.randn(5)
hess = hessian(f)(x)

Tracing through the transformations

We can also trace through these transformations in order to capture the results as new code using make_fx. There is also experimental integration with the NNC compiler (only works on CPU for now!).

from functorch import make_fx, grad
def f(x):
    return torch.sin(x).sum()
x = torch.randn(100)
grad_f = make_fx(grad(f))(x)
print(grad_f.code)

def forward(self, x_1):
    sin = torch.ops.aten.sin(x_1)
    sum_1 = torch.ops.aten.sum(sin, None);  sin = None
    cos = torch.ops.aten.cos(x_1);  x_1 = None
    _tensor_constant0 = self._tensor_constant0
    mul = torch.ops.aten.mul(_tensor_constant0, cos);  _tensor_constant0 = cos = None
    return mul

Working with NN modules: make_functional and friends

Sometimes you may want to perform a transform with respect to the parameters and/or buffers of an nn.Module. This can happen for example in:

  • model ensembling, where all of your weights and buffers have an additional dimension
  • per-sample-gradient computation where you want to compute per-sample-grads of the loss with respect to the model parameters

Our solution to this right now is an API that, given an nn.Module, creates a stateless version of it that can be called like a function.

  • make_functional(model) returns a functional version of model and the model.parameters()
  • make_functional_with_buffers(model) returns a functional version of model and the model.parameters() and model.buffers().

Here's an example where we compute per-sample-gradients using an nn.Linear layer:

import torch
from functorch import make_functional, vmap, grad

model = torch.nn.Linear(3, 3)
data = torch.randn(64, 3)
targets = torch.randn(64, 3)

func_model, params = make_functional(model)

def compute_loss(params, data, targets):
    preds = func_model(params, data)
    return torch.mean((preds - targets) ** 2)

per_sample_grads = vmap(grad(compute_loss), (None, 0, 0))(params, data, targets)

If you're making an ensemble of models, you may find combine_state_for_ensemble useful.

Documentation

For more documentation, see our docs website.

Debugging

torch._C._functorch.dump_tensor: Dumps dispatch keys on stack torch._C._functorch._set_vmap_fallback_warning_enabled(False) if the vmap warning spam bothers you.

Future Plans

In the end state, we'd like to upstream this into PyTorch once we iron out the design details. To figure out the details, we need your help -- please send us your use cases by starting a conversation in the issue tracker or trying our project out.

License

Functorch has a BSD-style license, as found in the LICENSE file.

Citing functorch

If you use functorch in your publication, please cite it by using the following BibTeX entry.

@Misc{functorch2021,
  author =       {Horace He, Richard Zou},
  title =        {functorch: JAX-like composable function transforms for PyTorch},
  howpublished = {\url{https://github.com/pytorch/functorch}},
  year =         {2021}
}

More Repositories

1

pytorch

Tensors and Dynamic neural networks in Python with strong GPU acceleration
Python
83,553
star
2

examples

A set of examples around pytorch in Vision, Text, Reinforcement Learning, etc.
Python
22,311
star
3

vision

Datasets, Transforms and Models specific to Computer Vision
Python
15,925
star
4

tutorials

PyTorch tutorials.
Jupyter Notebook
8,075
star
5

captum

Model interpretability and understanding for PyTorch
Python
4,781
star
6

ignite

High-level library to help with training and evaluating neural networks in PyTorch flexibly and transparently.
Python
4,507
star
7

serve

Serve, optimize and scale PyTorch models in production
Java
4,190
star
8

torchtune

PyTorch native finetuning library
Python
4,163
star
9

text

Models, data loaders and abstractions for language processing, powered by PyTorch
Python
3,490
star
10

ELF

ELF: a platform for game research with AlphaGoZero/AlphaZero reimplementation
C++
3,364
star
11

glow

Compiler for Neural Network hardware accelerators
C++
3,197
star
12

botorch

Bayesian optimization in PyTorch
Jupyter Notebook
3,043
star
13

torchchat

Run PyTorch LLMs locally on servers, desktop and mobile
Python
3,040
star
14

TensorRT

PyTorch/TorchScript/FX compiler for NVIDIA GPUs using TensorRT
Python
2,565
star
15

audio

Data manipulation and transformation for audio signal processing, powered by PyTorch
Python
2,471
star
16

xla

Enabling PyTorch on XLA Devices (e.g. Google TPU)
C++
2,469
star
17

rl

A modular, primitive-first, python-first PyTorch library for Reinforcement Learning.
Python
2,241
star
18

torchtitan

A native PyTorch Library for large model training
Python
2,130
star
19

executorch

On-device AI across mobile, embedded and edge for PyTorch
C++
1,954
star
20

torchrec

Pytorch domain library for recommendation systems
Python
1,852
star
21

opacus

Training PyTorch models with differential privacy
Jupyter Notebook
1,666
star
22

tnt

A lightweight library for PyTorch training tools and utilities
Python
1,650
star
23

QNNPACK

Quantized Neural Network PACKage - mobile-optimized implementation of quantized neural network operators
C
1,519
star
24

android-demo-app

PyTorch android examples of usage in applications
Java
1,460
star
25

hub

Submission to https://pytorch.org/hub/
Python
1,384
star
26

FBGEMM

FB (Facebook) + GEMM (General Matrix-Matrix Multiplication) - https://code.fb.com/ml-applications/fbgemm/
C++
1,156
star
27

data

A PyTorch repo for data loading and utilities to be shared by the PyTorch domain libraries.
Python
1,112
star
28

cpuinfo

CPU INFOrmation library (x86/x86-64/ARM/ARM64, Linux/Windows/Android/macOS/iOS)
C
989
star
29

torchdynamo

A Python-level JIT compiler designed to make unmodified PyTorch programs faster.
Python
989
star
30

extension-cpp

C++ extensions in PyTorch
Python
980
star
31

benchmark

TorchBench is a collection of open source benchmarks used to evaluate PyTorch performance.
Python
841
star
32

translate

Translate - a PyTorch Language Library
Python
820
star
33

tensordict

TensorDict is a pytorch dedicated tensor container.
Python
816
star
34

elastic

PyTorch elastic training
Python
728
star
35

PiPPy

Pipeline Parallelism for PyTorch
Python
698
star
36

kineto

A CPU+GPU Profiling library that provides access to timeline traces and hardware performance counters.
HTML
682
star
37

torcharrow

High performance model preprocessing library on PyTorch
Python
641
star
38

ao

PyTorch native quantization and sparsity for training and inference
Python
630
star
39

ios-demo-app

PyTorch iOS examples
Swift
595
star
40

tvm

TVM integration into PyTorch
C++
451
star
41

contrib

Implementations of ideas from recent papers
Python
390
star
42

ort

Accelerate PyTorch models with ONNX Runtime
Python
353
star
43

builder

Continuous builder and binary build scripts for pytorch
Shell
325
star
44

torchx

TorchX is a universal job launcher for PyTorch applications. TorchX is designed to have fast iteration time for training/research and support for E2E production ML pipelines when you're ready.
Python
319
star
45

accimage

high performance image loading and augmenting routines mimicking PIL.Image interface
C
317
star
46

extension-ffi

Examples of C extensions for PyTorch
Python
257
star
47

nestedtensor

[Prototype] Tools for the concurrent manipulation of variably sized Tensors.
Jupyter Notebook
252
star
48

tensorpipe

A tensor-aware point-to-point communication primitive for machine learning
C++
247
star
49

pytorch.github.io

The website for PyTorch
HTML
222
star
50

torcheval

A library that contains a rich collection of performant PyTorch model metrics, a simple interface to create new metrics, a toolkit to facilitate metric computation in distributed training and tools for PyTorch model evaluations.
Python
210
star
51

cppdocs

PyTorch C++ API Documentation
HTML
206
star
52

workshops

This is a repository for all workshop related materials.
Jupyter Notebook
204
star
53

hydra-torch

Configuration classes enabling type-safe PyTorch configuration for Hydra apps
Python
199
star
54

multipy

torch::deploy (multipy for non-torch uses) is a system that lets you get around the GIL problem by running multiple Python interpreters in a single C++ process.
C++
169
star
55

torchsnapshot

A performant, memory-efficient checkpointing library for PyTorch applications, designed with large, complex distributed workloads in mind.
Python
142
star
56

java-demo

Jupyter Notebook
126
star
57

rfcs

PyTorch RFCs (experimental)
120
star
58

torchdistx

Torch Distributed Experimental
Python
115
star
59

extension-script

Example repository for custom C++/CUDA operators for TorchScript
Python
112
star
60

csprng

Cryptographically secure pseudorandom number generators for PyTorch
Batchfile
105
star
61

pytorch_sphinx_theme

PyTorch Sphinx Theme
CSS
94
star
62

test-infra

This repository hosts code that supports the testing infrastructure for the main PyTorch repo. For example, this repo hosts the logic to track disabled tests and slow tests, as well as our continuation integration jobs HUD/dashboard.
TypeScript
78
star
63

expecttest

Python
71
star
64

torchcodec

PyTorch video decoding
Python
46
star
65

maskedtensor

MaskedTensors for PyTorch
Python
38
star
66

add-annotations-github-action

A GitHub action to run clang-tidy and annotate failures
JavaScript
13
star
67

ci-hud

HUD for CI activity on `pytorch/pytorch`, provides a top level view for jobs to easily discern regressions
JavaScript
11
star
68

probot

PyTorch GitHub bot written in probot
TypeScript
11
star
69

ossci-job-dsl

Jenkins job definitions for OSSCI
Groovy
10
star
70

pytorch-integration-testing

Testing downstream libraries using pytorch release candidates
Makefile
6
star
71

docs

This repository is automatically generated to contain the website source for the PyTorch documentation at https//pytorch.org/docs.
HTML
4
star
72

torchhub_testing

Repo to test torchhub. Nothing to see here.
4
star
73

dr-ci

Diagnose and remediate CI jobs
Haskell
2
star
74

pytorch-ci-dockerfiles

Scripts for generating docker images for PyTorch CI
2
star
75

labeler-github-action

GitHub action for labeling issues and pull requests based on conditions
TypeScript
1
star