Introducing Lightning Flash- the fastest way to get a DL baseline

[edenlightning]

Lightning Flash is a framework of tasks for fast prototyping, baselining, finetuning and solving business and scientific problems with deep learning. It is built for data scientists, machine learning practitioners, and applied researchers.

It is focused on:

• rapid baselining
• solving simple/boilerplate type of problems
• using pretrained models for finetuning and predictions.

It’s built on top of Lightning, so Flash can scale up across any hardware (GPUs, TPUS) with zero changes to your code.
Flash makes the power of Lightning more accessible to data scientists, developers and Kagglers, and makes baselining trivial for more experienced researchers. Similar to Lightning, everything is extremely modular so you can override any part of Flash with Lightning for full flexibility!

Read more here

import flash
from flash import download_data
from flash.vision import ImageClassificationData, ImageClassifier
from flash.core.finetuning import FreezeUnfreeze


# 1. Download the data
download_data("https://pl-flash-data.s3.amazonaws.com/hymenoptera_data.zip", 'data/')

# 2. Load the data
datamodule = ImageClassificationData.from_folders(
    backbone="resnet18",
    train_folder="data/hymenoptera_data/train/",
    valid_folder="data/hymenoptera_data/val/",
    test_folder="data/hymenoptera_data/test/",
)

# 3. Build the model
model = ImageClassifier(num_classes=datamodule.num_classes)

# 4. Create the trainer
trainer = flash.Trainer(max_epochs=5)

# 5. Finetune the model
trainer.finetune(model, datamodule=datamodule, strategy="freeze_unfreeze")

# 6. Use model for inference
predictions = model.predict([
    "data/hymenoptera_data/val/bees/65038344_52a45d090d.jpg",
    "data/hymenoptera_data/val/bees/590318879_68cf112861.jpg",
    "data/hymenoptera_data/val/ants/540543309_ddbb193ee5.jpg",
])
# Expect [1,1,0]
print(predictions)

We’re open to new task contributions! DM me or join our slack to learn more.