Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/szymonmaszke/torchfunc
PyTorch functions and utilities to make your life easier
https://github.com/szymonmaszke/torchfunc
docker extensions functions neural-network performance performance-analysis pytorch record recording tips utilities utils
Last synced: about 2 months ago
JSON representation
PyTorch functions and utilities to make your life easier
- Host: GitHub
- URL: https://github.com/szymonmaszke/torchfunc
- Owner: szymonmaszke
- License: mit
- Created: 2019-09-16T00:25:53.000Z (about 5 years ago)
- Default Branch: master
- Last Pushed: 2021-03-18T10:51:48.000Z (over 3 years ago)
- Last Synced: 2024-04-27T08:20:54.253Z (5 months ago)
- Topics: docker, extensions, functions, neural-network, performance, performance-analysis, pytorch, record, recording, tips, utilities, utils
- Language: Python
- Homepage:
- Size: 1.55 MB
- Stars: 195
- Watchers: 3
- Forks: 9
- Open Issues: 2
-
Metadata Files:
- Readme: README.md
- License: LICENSE
- Code of conduct: CODE_OF_CONDUCT.md
Awesome Lists containing this project
README
* Improve and analyse performance of your neural network (e.g. Tensor Cores compatibility)
* Record/analyse internal state of `torch.nn.Module` as data passes through it
* Do the above based on external conditions (using single `Callable` to specify it)
* Day-to-day neural network related duties (model size, seeding, time measurements etc.)
* Get information about your host operating system, `torch.nn.Module` device, CUDA
capabilities etc.
| Version | Docs | Tests | Coverage | Style | PyPI | Python | PyTorch | Docker | Roadmap |
|---------|------|-------|----------|-------|------|--------|---------|--------|---------|
| [](https://github.com/szymonmaszke/torchfunc/releases) | [](https://szymonmaszke.github.io/torchfunc/) |  |  | [](https://codebeat.co/projects/github-com-szymonmaszke-torchfunc-master) | [](https://pypi.org/project/torchfunc/) | [](https://www.python.org/) | [](https://pytorch.org/) | [](https://hub.docker.com/r/szymonmaszke/torchfunc) | [](https://github.com/szymonmaszke/torchfunc/blob/master/ROADMAP.md) |
# :bulb: Examples
__Check documentation here:__ [https://szymonmaszke.github.io/torchfunc](https://szymonmaszke.github.io/torchfunc)
## 1. Getting performance tips
- __Get instant performance tips about your module. All problems described by comments
will be shown by `torchfunc.performance.tips`:__
```python
class Model(torch.nn.Module):
def __init__(self):
super().__init__()
self.convolution = torch.nn.Sequential(
torch.nn.Conv2d(1, 32, 3),
torch.nn.ReLU(inplace=True), # Inplace may harm kernel fusion
torch.nn.Conv2d(32, 128, 3, groups=32), # Depthwise is slower in PyTorch
torch.nn.ReLU(inplace=True), # Same as before
torch.nn.Conv2d(128, 250, 3), # Wrong output size for TensorCores
)
self.classifier = torch.nn.Sequential(
torch.nn.Linear(250, 64), # Wrong input size for TensorCores
torch.nn.ReLU(), # Fine, no info about this layer
torch.nn.Linear(64, 10), # Wrong output size for TensorCores
)
def forward(self, inputs):
convolved = torch.nn.AdaptiveAvgPool2d(1)(self.convolution(inputs)).flatten()
return self.classifier(convolved)
# All you have to do
print(torchfunc.performance.tips(Model()))
```
## 2. Seeding, weight freezing and others
- __Seed globaly (including `numpy` and `cuda`), freeze weights, check inference time and model size:__
```python
# Inb4 MNIST, you can use any module with those functions
model = torch.nn.Linear(784, 10)
torchfunc.seed(0)
frozen = torchfunc.module.freeze(model, bias=False)
with torchfunc.Timer() as timer:
frozen(torch.randn(32, 784)
print(timer.checkpoint()) # Time since the beginning
frozen(torch.randn(128, 784)
print(timer.checkpoint()) # Since last checkpoint
print(f"Overall time {timer}; Model size: {torchfunc.sizeof(frozen)}")
```
## 3. Record `torch.nn.Module` internal state
- __Record and sum per-layer activation statistics as data passes through network:__
```python
# Still MNIST but any module can be put in it's place
model = torch.nn.Sequential(
torch.nn.Linear(784, 100),
torch.nn.ReLU(),
torch.nn.Linear(100, 50),
torch.nn.ReLU(),
torch.nn.Linear(50, 10),
)
# Recorder which sums all inputs to layers
recorder = torchfunc.hooks.recorders.ForwardPre(reduction=lambda x, y: x+y)
# Record only for torch.nn.Linear
recorder.children(model, types=(torch.nn.Linear,))
# Train your network normally (or pass data through it)
...
# Activations of all neurons of first layer!
print(recorder[1]) # You can also post-process this data easily with apply
```
For other examples (and how to use condition), see [documentation](https://szymonmaszke.github.io/torchfunc/)
# :wrench: Installation
## :snake: [pip]()
### Latest release:
```shell
pip install --user torchfunc
```
### Nightly:
```shell
pip install --user torchfunc-nightly
```
## :whale2: [Docker](https://hub.docker.com/r/szymonmaszke/torchfunc)
__CPU standalone__ and various versions of __GPU enabled__ images are available
at [dockerhub](https://hub.docker.com/r/szymonmaszke/torchfunc/tags).
For CPU quickstart, issue:
```shell
docker pull szymonmaszke/torchfunc:18.04
```
Nightly builds are also available, just prefix tag with `nightly_`. If you are going for `GPU` image make sure you have
[nvidia/docker](https://github.com/NVIDIA/nvidia-docker) installed and it's runtime set.
# :question: Contributing
If you find any issue or you think some functionality may be useful to others and fits this library, please [open new Issue](https://help.github.com/en/articles/creating-an-issue) or [create Pull Request](https://help.github.com/en/articles/creating-a-pull-request-from-a-fork).
To get an overview of things one can do to help this project, see [Roadmap](https://github.com/szymonmaszke/torchfunc/blob/master/ROADMAP.md).