efficientnet

What it is and what it does

EfficientNet is a family of convolutional neural network architectures (B0 through B7) designed to achieve high ImageNet accuracy with significantly fewer parameters and computational cost than competing models. This package provides Keras and TensorFlow Keras implementations with pre-trained weights, allowing you to load a model and use it immediately for image classification, transfer learning, or as a backbone for custom tasks.

The package depends on keras-applications for model utilities and scikit-image for image processing. It supports both standalone Keras and TensorFlow's integrated Keras API. Models can be initialized with ImageNet weights or noisy-student weights; the description notes a major update on 24 July 2019 that unified support across both frameworks, though models trained before that date require version 0.0.4.

Use it for:

• Load a pre-trained EfficientNet model for immediate image classification on new images without retraining.
• Use EfficientNet as a feature extractor backbone for transfer learning on custom image datasets.
• Compare model variants (B0–B7) to balance accuracy and inference speed for deployment constraints.
• Fine-tune a pre-trained EfficientNet on domain-specific images with limited labeled data.
• Benchmark model efficiency: parameter count and FLOPS against other architectures like ResNet.

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Provides Keras and TensorFlow Keras implementations of EfficientNet, a family of lightweight convolutional neural networks (B0–B7) with pre-trained ImageNet weights for transfer learning and inference.

Yes, if you need a pre-trained EfficientNet for Keras/TensorFlow and are working with compatible framework versions. The low install friction and permissive license make it straightforward to add. However, the package is dormant (last release September 2020, last commit January 2024), so verify compatibility with your current Keras/TensorFlow versions before committing to production use. For active maintenance and broader framework support, consider alternatives that are actively maintained.

Install

efficientnet on PyPI

pip install efficientnet

uv add efficientnet

poetry add efficientnet

Installing efficientnet

Before you install

Low install friction with only two runtime dependencies (keras-applications and scikit-image). However, the package is dormant—last release was 2020-09-15 and last commit 2024-01-24—so it may not receive updates for newer Keras or TensorFlow versions.

License in practice

Licensed under Apache License 2.0 (permissive), allowing commercial use, modification, and distribution with minimal restrictions.

Quickstart

pip install efficientnet

import efficientnet.keras as efn
model = efn.EfficientNetB0(weights='imagenet')

Requires Keras >= 2.2.0 or TensorFlow >= 1.12.0 to be installed separately; the package itself does not declare these as direct dependencies.

Verify before relying

• Whether the package works with current TensorFlow/Keras versions (last release 2020-09-15, last commit 2024-01-24).
• Whether pre-trained weights are still accessible and up-to-date.
• Compatibility with modern Python versions beyond what 'supports_current' indicates.

Package facts

Evidence: efficientnet-1.1.1-py3-none-any.whl

Tags

More Artificial Intelligence packages