Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/lightning-universe/lightning-flash
Your PyTorch AI Factory - Flash enables you to easily configure and run complex AI recipes for over 15 tasks across 7 data domains
https://github.com/lightning-universe/lightning-flash
classification deep-learning fiftyone icevision machine-learning object-detection open3d pytorch pytorch-lightning pytorch-video tabular-data tasks-flash torch-geometric
Last synced: about 2 months ago
JSON representation
Your PyTorch AI Factory - Flash enables you to easily configure and run complex AI recipes for over 15 tasks across 7 data domains
- Host: GitHub
- URL: https://github.com/lightning-universe/lightning-flash
- Owner: Lightning-Universe
- License: apache-2.0
- Archived: true
- Created: 2021-01-28T18:47:16.000Z (almost 4 years ago)
- Default Branch: master
- Last Pushed: 2023-10-08T14:28:44.000Z (about 1 year ago)
- Last Synced: 2024-05-22T08:11:25.962Z (6 months ago)
- Topics: classification, deep-learning, fiftyone, icevision, machine-learning, object-detection, open3d, pytorch, pytorch-lightning, pytorch-video, tabular-data, tasks-flash, torch-geometric
- Language: Python
- Homepage: https://lightning-flash.readthedocs.io
- Size: 12.8 MB
- Stars: 1,729
- Watchers: 36
- Forks: 212
- Open Issues: 30
-
Metadata Files:
- Readme: README.md
- Changelog: CHANGELOG.md
- Contributing: .github/CONTRIBUTING.md
- License: LICENSE
- Codeowners: .github/CODEOWNERS
Awesome Lists containing this project
README
**Your PyTorch AI Factory**
---
Installation •
Flash in 3 Steps •
Docs •
Contribute •
Community •
Website •
License
[](https://pypi.org/project/lightning-flash/)
[](https://badge.fury.io/py/lightning-flash)
[](https://www.pepy.tech/projects/lightning-flash)
[](https://www.pytorchlightning.ai/community)
[](https://github.com/Lightning-AI/pytorch-lightning/blob/master/LICENSE)
[](https://github.com/Lightning-Universe/lightning-flash/actions/workflows/ci-testing.yml)
[](https://codecov.io/gh/Lightning-Universe/lightning-flash)
[](https://lightning-flash.readthedocs.io/en/stable/?badge=stable)
[](https://zenodo.org/badge/latestdoi/333857397)
---
Flash makes complex AI recipes for over 15 tasks across 7 data domains accessible to all.
In a nutshell, Flash is the production grade research framework you always dreamed of but didn't have time to build.
## Getting Started
From PyPI:
```bash
pip install lightning-flash
```
See [our installation guide](https://lightning-flash.readthedocs.io/en/latest/installation.html) for more options.
## Flash in 3 Steps
### Step 1. Load your data
All data loading in Flash is performed via a `from_*` classmethod on a `DataModule`.
To decide which `DataModule` to use and which `from_*` methods are available, it depends on the task you want to perform.
For example, for image segmentation where your data is stored in folders, you would use the [`from_folders` method of the `SemanticSegmentationData` class](https://lightning-flash.readthedocs.io/en/latest/reference/semantic_segmentation.html#from-folders):
```py
from flash.image import SemanticSegmentationData
dm = SemanticSegmentationData.from_folders(
train_folder="data/CameraRGB",
train_target_folder="data/CameraSeg",
val_split=0.1,
image_size=(256, 256),
num_classes=21,
)
```
### Step 2: Configure your model
Our tasks come loaded with pre-trained backbones and (where applicable) heads.
You can view the available backbones to use with your task using [`available_backbones`](https://lightning-flash.readthedocs.io/en/latest/general/backbones.html).
Once you've chosen one, create the model:
```py
from flash.image import SemanticSegmentation
print(SemanticSegmentation.available_heads())
# ['deeplabv3', 'deeplabv3plus', 'fpn', ..., 'unetplusplus']
print(SemanticSegmentation.available_backbones('fpn'))
# ['densenet121', ..., 'xception'] # + 113 models
print(SemanticSegmentation.available_pretrained_weights('efficientnet-b0'))
# ['imagenet', 'advprop']
model = SemanticSegmentation(
head="fpn", backbone='efficientnet-b0', pretrained="advprop", num_classes=dm.num_classes)
```
### Step 3: Finetune!
```py
from flash import Trainer
trainer = Trainer(max_epochs=3)
trainer.finetune(model, datamodule=datamodule, strategy="freeze")
trainer.save_checkpoint("semantic_segmentation_model.pt")
```
---
## PyTorch Recipes
### Make predictions with Flash!
Serve in just 2 lines:
```py
from flash.image import SemanticSegmentation
model = SemanticSegmentation.load_from_checkpoint("semantic_segmentation_model.pt")
model.serve()
```
or make predictions from raw data directly.
```py
from flash import Trainer
trainer = Trainer(strategy='ddp', accelerator="gpu", gpus=2)
dm = SemanticSegmentationData.from_folders(predict_folder="data/CameraRGB")
predictions = trainer.predict(model, dm)
```
### Flash Training Strategies
Training strategies are PyTorch SOTA Training Recipes which can be utilized with a given task.
Check out this [example](https://github.com/Lightning-AI/lightning-flash/blob/master/examples/integrations/learn2learn/image_classification_imagenette_mini.py) where the `ImageClassifier` supports 4 [Meta Learning Algorithms](https://lilianweng.github.io/lil-log/2018/11/30/meta-learning.html) from [Learn2Learn](https://github.com/learnables/learn2learn).
This is particularly useful if you use this model in production and want to make sure the model adapts quickly to its new environment with minimal labelled data.
```py
from flash.image import ImageClassifier
model = ImageClassifier(
backbone="resnet18",
optimizer=torch.optim.Adam,
optimizer_kwargs={"lr": 0.001},
training_strategy="prototypicalnetworks",
training_strategy_kwargs={
"epoch_length": 10 * 16,
"meta_batch_size": 4,
"num_tasks": 200,
"test_num_tasks": 2000,
"ways": datamodule.num_classes,
"shots": 1,
"test_ways": 5,
"test_shots": 1,
"test_queries": 15,
},
)
```
In detail, the following methods are currently implemented:
* **[prototypicalnetworks](https://github.com/learnables/learn2learn/blob/master/learn2learn/algorithms/lightning/lightning_protonet.py)** : from Snell *et al.* 2017, [Prototypical Networks for Few-shot Learning](https://arxiv.org/abs/1703.05175)
* **[maml](https://github.com/learnables/learn2learn/blob/master/learn2learn/algorithms/lightning/lightning_maml.py)** : from Finn *et al.* 2017, [Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks](https://arxiv.org/abs/1703.03400)
* **[metaoptnet](https://github.com/learnables/learn2learn/blob/master/learn2learn/algorithms/lightning/lightning_metaoptnet.py)** : from Lee *et al.* 2019, [Meta-Learning with Differentiable Convex Optimization](https://arxiv.org/abs/1904.03758)
* **[anil](https://github.com/learnables/learn2learn/blob/master/learn2learn/algorithms/lightning/lightning_anil.py)** : from Raghu *et al.* 2020, [Rapid Learning or Feature Reuse? Towards Understanding the Effectiveness of MAML](https://arxiv.org/abs/1909.09157)
### Flash Optimizers / Schedulers
With Flash, swapping among 40+ optimizers and 15+ schedulers recipes are simple. Find the list of available optimizers, schedulers as follows:
```py
from flash.image import ImageClassifier
ImageClassifier.available_optimizers()
# ['A2GradExp', ..., 'Yogi']
ImageClassifier.available_schedulers()
# ['CosineAnnealingLR', 'CosineAnnealingWarmRestarts', ..., 'polynomial_decay_schedule_with_warmup']
```
Once you've chosen, create the model:
```py
#### The optimizer of choice can be passed as
from flash.image import ImageClassifier
# - String value
model = ImageClassifier(backbone="resnet18", num_classes=2, optimizer="Adam", lr_scheduler=None)
# - Callable
model = ImageClassifier(backbone="resnet18", num_classes=2, optimizer=functools.partial(torch.optim.Adadelta, eps=0.5), lr_scheduler=None)
# - Tuple[string, dict]: (The dict takes in the optimizer kwargs)
model = ImageClassifier(backbone="resnet18", num_classes=2, optimizer=("Adadelta", {"epa": 0.5}), lr_scheduler=None)
#### The scheduler of choice can be passed as a
# - String value
model = ImageClassifier(backbone="resnet18", num_classes=2, optimizer="Adam", lr_scheduler="constant_schedule")
# - Callable
model = ImageClassifier(backbone="resnet18", num_classes=2, optimizer="Adam", lr_scheduler=functools.partial(CyclicLR, step_size_up=1500, mode='exp_range', gamma=0.5))
# - Tuple[string, dict]: (The dict takes in the scheduler kwargs)
model = ImageClassifier(backbone="resnet18", num_classes=2, optimizer="Adam", lr_scheduler=("StepLR", {"step_size": 10}))
```
You can also register you own custom scheduler recipes beforeahand and use them shown as above:
```py
from flash.image import ImageClassifier
@ImageClassifier.lr_schedulers_registry
def my_steplr_recipe(optimizer):
return torch.optim.lr_scheduler.StepLR(optimizer, step_size=10)
model = ImageClassifier(backbone="resnet18", num_classes=2, optimizer="Adam", lr_scheduler="my_steplr_recipe")
```
### Flash Transforms
Flash includes some simple augmentations for each task by default, however, you will often want to override these and control your own augmentation recipe.
To this end, Flash supports custom transformations with the [`InputTransform`](https://lightning-flash.readthedocs.io/en/stable/api/generated/flash.core.data.io.input_transform.InputTransform.html).
The `InputTransform` is like a callback for transforms, with hooks that can be used to apply transforms to samples or batches, on and off the device / accelerator.
In addition, hooks can be specialized to apply transforms only to the input or target.
With these hooks, complex transforms like MixUp can be implemented with ease.
Here's an example (with an albumentations transform thrown in too!):
```py
import torch
import numpy as np
import albumentations
from flash import InputTransform
from flash.image import ImageClassificationData
from flash.image.classification.input_transform import AlbumentationsAdapter
def mixup(batch, alpha=1.0):
images = batch["input"]
targets = batch["target"].float().unsqueeze(1)
lam = np.random.beta(alpha, alpha)
perm = torch.randperm(images.size(0))
batch["input"] = images * lam + images[perm] * (1 - lam)
batch["target"] = targets * lam + targets[perm] * (1 - lam)
return batch
class MixUpInputTransform(InputTransform):
def train_input_per_sample_transform(self):
return AlbumentationsAdapter(albumentations.HorizontalFlip(p=0.5))
# This will be applied after transferring the batch to the device!
def train_per_batch_transform_on_device(self):
return mixup
datamodule = ImageClassificationData.from_folders(
train_folder="data/train",
transform=MixUpInputTransform,
batch_size=2,
)
```
## Flash Zero - PyTorch Recipes from the Command Line!
Flash Zero is a zero-code machine learning platform built
directly into lightning-flash
using the [`Lightning CLI`](https://pytorch-lightning.readthedocs.io/en/stable/common/lightning_cli.html).
To get started and view the available tasks, run:
```bash
flash --help
```
For example, to train an image classifier for 10 epochs with a `resnet50` backbone on 2 GPUs using your own data, you can do:
```bash
flash image_classification --trainer.max_epochs 10 --trainer.gpus 2 --model.backbone resnet50 from_folders --train_folder {PATH_TO_DATA}
```
## Kaggle Notebook Examples
- [🚢Titanic crash with Lightning⚡Flash](https://www.kaggle.com/jirkaborovec/titanic-crash-with-lightning-flash)
- [🏠House 💵prices predictions with Lightning⚡Flash](https://www.kaggle.com/jirkaborovec/house-prices-predictions-with-lightning-flash)
- [Playing 📋tabular with Lightning⚡Flash](https://www.kaggle.com/jirkaborovec/playing-tabular-with-lightning-flash)
- [🙊Toxic comments with Lightning⚡Flash](https://www.kaggle.com/jirkaborovec/toxic-comments-with-lightning-flash)
- [🫁COVID detection with Lightning⚡️Flash](https://www.kaggle.com/jirkaborovec/covid-detection-with-lightning-flash)
## Contribute!
The lightning + Flash team is hard at work building more tasks for common deep-learning use cases. But we're looking for incredible contributors like you to submit new tasks!
Join our [Slack](https://www.pytorchlightning.ai/community) and/or read our [CONTRIBUTING](https://github.com/PyTorchLightning/lightning-flash/blob/master/.github/CONTRIBUTING.md) guidelines to get help becoming a contributor!
__Note:__ Flash is currently being tested on real-world use cases and is in active development. Please [open an issue](https://github.com/PyTorchLightning/lightning-flash/issues/new/choose) if you find anything that isn't working as expected.
---
## Community
Flash is maintained by our [core contributors](https://lightning-flash.readthedocs.io/en/latest/governance.html).
For help or questions, join our huge community on [Slack](https://www.pytorchlightning.ai/community)!
---
## Citations
We’re excited to continue the strong legacy of opensource software and have been inspired over the years by Caffe, Theano, Keras, PyTorch, torchbearer, and [fast.ai](https://arxiv.org/abs/2002.04688). When/if additional papers are written about this, we’ll be happy to cite these frameworks and the corresponding authors.
Flash leverages models from many different frameworks in order to cover such a wide range of domains and tasks. The full list of providers can be found in [our documentation](https://lightning-flash.readthedocs.io/en/latest/integrations/providers.html).
---
## License
Please observe the Apache 2.0 license that is listed in this repository.