๐ Quick Start#
Access our library of speedup methods with the ฦ() Functional API:
import logging
from composer import functional as CF
import torchvision.models as models
logging.basicConfig(level=logging.INFO)
model = models.resnet50()
CF.apply_blurpool(model)
This creates a ResNet50 model and replaces several pooling and convolution layers with BlurPool variants (Zhang et al, 2019). For more information, see ๐ BlurPool. The method should log:
Applied BlurPool to model ResNet. Model now has 1 BlurMaxPool2d and 6 BlurConv2D layers.
These methods are easy to integrate into your own training loop code with just a few lines.
For an overview of the algorithms, see ๐ค Algorithms.
We make composing recipes together even easier with our (optional) Trainer
. Here,
we train an MNIST classifer with a recipe of methods:
from torchvision import datasets, transforms
from torch.utils.data import DataLoader
from composer import Trainer
from composer.models import mnist_model
from composer.algorithms import LabelSmoothing, CutMix, ChannelsLast
transform = transforms.Compose([transforms.ToTensor()])
dataset = datasets.MNIST("data", train=True, download=True, transform=transform)
train_dataloader = DataLoader(dataset, batch_size=128)
trainer = Trainer(
model=mnist_model(num_classes=10),
train_dataloader=train_dataloader,
max_duration="2ep",
algorithms=[
LabelSmoothing(smoothing=0.1),
CutMix(alpha=1.0),
ChannelsLast(),
]
)
trainer.fit()
We handle inserting and running the logic during the training so that any algorithms you specify โjust work.โ
Besides easily running our built-in algorithms, Composer also features:
An interface to flexibly add algorithms to the training loop
An engine that manages the ordering of algorithms for composition
A trainer to handle boilerplate around numerics, distributed training, and others
Integration with popular model libraries such as HuggingFace Transformers
Next steps#
Try our ๐ผ๏ธ Getting Started tutorial on Colab.
See โ๏ธ Using the Trainer for more details on our trainer.
Read ๐ Welcome Tour for a tour through the library.