---
license: apache-2.0
library_name: litert
pipeline_tag: image-classification
datasets:
- imagenet-1k
tags:
- mediapipe
- litert
- tflite
- image-classification
- on-device
- android
- ios
- web
- python
- flutter
- vision
- efficientnet
- efficientnet-lite2
- int8
- fp32
- arxiv:1905.11946
---
# MediaPipe EfficientNet-Lite2 (Image Classifier)
EfficientNet-Lite2 1,000-class ImageNet image classification models in full-integer INT8 (`efficientnet_lite2_int8.tflite`, 7.2 MB) and FP32 (`efficientnet_lite2_float32.tflite`, 24.3 MB), offering higher accuracy than EfficientNet-Lite0 for the [MediaPipe Image Classifier (`ImageClassifier`)](https://developers.google.com/edge/mediapipe/solutions/vision/image_classifier) task and standalone [LiteRT](https://developers.google.com/edge/litert) inference.
EfficientNet-Lite2 scales up resolution (`[1, 260, 260, 3]` RGB input) and network capacity over EfficientNet-Lite0 while keeping ReLU6 activations and edge-friendly layers.
### Embedded vs. Standalone Label Files
Both `.tflite` files embed `labels_without_background.txt` (**1,000 classes**, `tench` through `toilet tissue`) inside `TFLITE_METADATA`. The companion [`labels.txt`](./labels.txt) contains **1,001 lines** (line 1 is `background`, lines 2–1001 are the 1,000 ImageNet classes).
## Files
| File in Repo Root | Upstream Google Path | Precision | Size | SHA-256 |
|---|---|---|---:|---|
| [`efficientnet_lite2_int8.tflite`](https://huggingface.co/litert-community/MediaPipe-ImageClassifier-EfficientNet-Lite2/blob/main/efficientnet_lite2_int8.tflite) | `image_classifier/efficientnet_lite2/int8/latest/efficientnet_lite2.tflite` | `int8` | 7.17 MB (7,165,461 B) | `7fc3a7dbc1cd2833b162a54577afb47ab531d9c4b799ef11d90a4c56fe4ed5ad` |
| [`efficientnet_lite2_float32.tflite`](https://huggingface.co/litert-community/MediaPipe-ImageClassifier-EfficientNet-Lite2/blob/main/efficientnet_lite2_float32.tflite) | `image_classifier/efficientnet_lite2/float32/latest/efficientnet_lite2.tflite` | `float32` | 24.33 MB (24,325,765 B) | `94b6c84a5d6de20f932558de213b50e0338dcc06ab83d8d64b12128a476b4b83` |
### Companion Files in Repo Root
| File | Role | Size | SHA-256 |
|---|---|---:|---|
| [`labels.txt`](./labels.txt) | Label map | 10.5 KB (10,484 B) | `536feacc519de3d418de26b2effb4d75694a8c4c0063e36499a46fa8061e2da9` |
### Download
```bash
pip install -U huggingface_hub
hf download litert-community/MediaPipe-ImageClassifier-EfficientNet-Lite2 efficientnet_lite2_int8.tflite --local-dir .
hf download litert-community/MediaPipe-ImageClassifier-EfficientNet-Lite2 efficientnet_lite2_float32.tflite --local-dir .
```
## Inputs, Outputs, and Model Specifications
### `efficientnet_lite2_int8.tflite`
- Embedded `TFLITE_METADATA`: name `EfficientNet-lite image classifier (quantized)` · version `1` · author `MediaPipe` · packed associated files: `labels_without_background.txt`
| Direction | Tensor | Shape | Type | Notes |
|---|---|---|---|---|
| Input | `images` | `[1, 260, 260, 3]` | `uint8` | image — Input image to be classified. The expected image is 260 x 260, with three channels (red, blue, and green) per pixel. Each element in the tensor is a value between min and max, where (per-channel) min is [0] and max is [255]. — Normalization: `mean=[127.0], std=[128.0]` — Quantization: `scale=0.0123429, zero_point=125` |
| Output | `Softmax` | `[1, 1000]` | `uint8` | probability — Probabilities of the 1000 labels respectively. — Quantization: `scale=0.00390625, zero_point=0` — Labels: `labels_without_background.txt` |
### `efficientnet_lite2_float32.tflite`
- Embedded `TFLITE_METADATA`: name `EfficientNet-lite image classifier` · version `1` · author `MediaPipe` · packed associated files: `labels_without_background.txt`
| Direction | Tensor | Shape | Type | Notes |
|---|---|---|---|---|
| Input | `images` | `[1, 260, 260, 3]` | `float32` | image — Input image to be classified. The expected image is 260 x 260, with three channels (red, blue, and green) per pixel. Each element in the tensor is a value between min and max, where (per-channel) min is [-0.9921875] and max is [1.0]. — Normalization: `mean=[127.0], std=[128.0]` |
| Output | `Softmax` | `[1, 1000]` | `float32` | probability — Probabilities of the 1000 labels respectively. — Labels: `labels_without_background.txt` |
## Use with MediaPipe Tasks
### Python Quickstart
```bash
pip install -U mediapipe huggingface_hub
```
```python
import mediapipe as mp
from huggingface_hub import hf_hub_download
model_path = hf_hub_download(
repo_id="litert-community/MediaPipe-ImageClassifier-EfficientNet-Lite2",
filename="efficientnet_lite2_int8.tflite",
)
BaseOptions = mp.tasks.BaseOptions
ImageClassifier = mp.tasks.vision.ImageClassifier
ImageClassifierOptions = mp.tasks.vision.ImageClassifierOptions
VisionRunningMode = mp.tasks.vision.RunningMode
options = ImageClassifierOptions(
base_options=BaseOptions(model_asset_path=model_path),
running_mode=VisionRunningMode.IMAGE,
max_results=5,
)
with ImageClassifier.create_from_options(options) as runner:
mp_image = mp.Image.create_from_file("input.jpg")
result = runner.classify(mp_image)
for category in result.classifications[0].categories:
print(f"{category.category_name}: {category.score:.3f}")
```
### Cross-Platform Guides & Samples
- **Task overview & options:** [Image Classifier documentation](https://developers.google.com/edge/mediapipe/solutions/vision/image_classifier)
- **Platform setup guides:** [Python](https://developers.google.com/edge/mediapipe/solutions/setup_python) · [Web / JavaScript](https://developers.google.com/edge/mediapipe/solutions/setup_web) · [Android (Kotlin / Java)](https://developers.google.com/edge/mediapipe/solutions/setup_android) · [iOS (Swift)](https://developers.google.com/edge/mediapipe/solutions/setup_ios) · [Flutter (Dart)](https://developers.google.com/edge/mediapipe/solutions/setup_flutter)
- **Runnable sample apps:** [Web Demo](https://google-ai-edge.github.io/mediapipe-samples-web/#/vision/image_classifier) · [Web Source](https://github.com/google-ai-edge/mediapipe-samples-web/blob/main/src/tasks/image-classifier.ts) · [Android](https://github.com/google-ai-edge/mediapipe-samples/tree/main/examples/image_classification/android) · [iOS](https://github.com/google-ai-edge/mediapipe-samples/tree/main/examples/image_classification/ios) · [Python Notebook](https://colab.research.google.com/github/google-ai-edge/mediapipe-samples/blob/main/examples/image_classification/python/image_classifier.ipynb) · [Flutter](https://github.com/google/flutter-mediapipe/tree/main/samples/vision/image_classifier)
## Latency Benchmarks
End-to-end MediaPipe Task pipeline average latency reported by the [Image Classifier documentation](https://developers.google.com/edge/mediapipe/solutions/vision/image_classifier#benchmarks):
| Model Variant | Device | CPU Latency | GPU Latency |
|---|---|---:|---:|
| EfficientNet-Lite2 (int8) | Pixel 6 | 19.43 ms | — |
| EfficientNet-Lite2 (float 32) | Pixel 6 | 44.17 ms | 22.20 ms |
## Intended Use, Limitations, and Fairness
- **Intended applications:** Higher-accuracy on-device classification of prominent everyday subjects across the 1,000 ImageNet-1K synsets where additional latency over EfficientNet-Lite0 is acceptable.
- **Limitations & considerations:**
- Closed-vocabulary ImageNet-1K classifier; assumes a centered dominant subject inside a `260 x 260` crop.
- Not intended for safety-critical, medical, or biometric decisions without domain-specific fine-tuning.
- Google does not publish a separate standalone fairness PDF for EfficientNet-Lite2 on the MediaPipe documentation page.
## Provenance
Every model binary in this repository is byte-identical (unmodified FlatBuffer weights and embedded `TFLITE_METADATA` bytes) to the official artifact served by Google from `storage.googleapis.com/mediapipe-models`.
| Repo File | Upstream URL | GCS Generation | Upstream `Last-Modified` | Pinned Bucket Version |
|---|---|---|---|---|
| `efficientnet_lite2_int8.tflite` | [`efficientnet_lite2.tflite`](https://storage.googleapis.com/mediapipe-models/image_classifier/efficientnet_lite2/int8/latest/efficientnet_lite2.tflite) | `1682480012152147` | Wed, 26 Apr 2023 03:33:32 GMT | `/1/` |
| `efficientnet_lite2_float32.tflite` | [`efficientnet_lite2.tflite`](https://storage.googleapis.com/mediapipe-models/image_classifier/efficientnet_lite2/float32/latest/efficientnet_lite2.tflite) | `1682480010617750` | Wed, 26 Apr 2023 03:33:30 GMT | `/1/` |
## License and Citation
- License: `apache-2.0` — Stated as Apache License, Version 2.0 (`http://www.apache.org/licenses/LICENSE-2.0`) in the embedded `TFLITE_METADATA` block of the `.tflite` model file(s).
### Citation
```bibtex
@inproceedings{tan2019efficientnet,
title = {EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks},
author = {Tan, Mingxing and Le, Quoc V.},
booktitle = {International Conference on Machine Learning (ICML)},
year = {2019},
url = {https://arxiv.org/abs/1905.11946}
}
```