EfficientNetLite (MediaPipe)
EfficientNet-Lite is a family of image classification models designed for mobile/IoT usage.EfficientNet-Lite is a family of lightweight mobile-friendly Convolutional Neural Networks(CNNs) designed for vision ML tasks.
Derived from the EfficientNet architecture, EfficientNet-Lite removes operations like squeeze-and-excite(SE) and replaces swish activations with RELU6 to be more mobile-friendly. Just like the EfficientNet architecture family, the EfficientNet-Lite models include multiple architectures of different sizes. The following models and input image sizes are supported:
Model | Input image size |
---|---|
efficientnet-lite0 | 224x224 |
efficientnet-lite2 | 260x260 |
efficientnet-lite4 | 300x300 |
This model page is for the efficientnet-lite[0,2,4] models with a classification head.
The EfficientNet-Lite models are intended for on-device use cases. Using the notebook, you can create a custom EfficientNet-Lite model, which can be deployed on-device (Android, iOS, Web, desktop, etc) using MediaPipe Tasks ImageClassifier. Use MediaPipe Studio to evaluate the model through interactive live demo.
The demo below is for internal testing purposes only. Output should not be saved or distributed. Please do not provide personally identifiable information or other data subject to regulatory requirements.
This model can be used in a notebook. Click Open notebook to use the model in Colab.
This model checkpoint was pre-trained on the ILSVRC-2012-CLS dataset for image classification(ImageNet).
Training images are resized/rescaled to the same resolution and normalized such that all color values fall within the range of [0, 1]. For details on the input image resolution for each model, refer to the table above.
Given an image, the model will output a vector of confidence scores for each support class (label) the model identifies.
For best results, use MediaPipe Tasks ImageClassifier to deploy the output TFLite model on-device as it ensures the same preprocessing logic between training and inference. Use MediaPipe Studio to evaluate the model through interactive live demo.
Resource ID | Release date | Release stage | Description |
---|---|---|---|
mediapipe/efficientnetlite0 | 2024-04-01 | General Availability | Fine tuning and ondevice serving |
La console Google Cloud n'a pas pu charger les sources JavaScript depuis www.gstatic.com.
Voici les raisons possibles :