---
pipeline_tag: image-classification
license: apache-2.0
base_model: timm/mobilenetv2_050.lamb_in1k
library_name: kerasformers
tags:
- keras
- kerasformers
- image-classification
- mobilenetv2
- backbone
- arxiv:1801.04381
- pytorch
- jax
- tf
---

## ***See [our collection](https://huggingface.co/collections/kerasformers/mobilenetv2-6a6bd9774cd9d82c8edf9442) for all versions of MobileNetV2.***

# Run MobileNetV2 with Keras 3: JAX, PyTorch, or TensorFlow

[![GitHub](https://img.shields.io/badge/GitHub-KerasFormers-black?logo=github)](https://github.com/IMvision12/KerasFormers) [![Docs](https://img.shields.io/badge/Docs-MobileNetV2-blue)](https://imvision12.github.io/KerasFormers/classification_backbones/) [![Collection](https://img.shields.io/badge/HF-MobileNetV2%20collection-yellow)](https://huggingface.co/collections/kerasformers/mobilenetv2-6a6bd9774cd9d82c8edf9442)

# kerasformers/mobilenetv2_050_lamb_in1k

Paper: [MobileNetV2: Inverted Residuals and Linear Bottlenecks (arXiv:1801.04381)](https://arxiv.org/abs/1801.04381) · [HF Papers](https://huggingface.co/papers/1801.04381)

MobileNetV2 uses inverted residuals and linear bottlenecks for mobile classification and as a lightweight multi-scale backbone.

For more details on the model, please go to the upstream [model card](https://huggingface.co/timm/mobilenetv2_050.lamb_in1k).

Pure-**Keras 3** conversion of [`timm/mobilenetv2_050.lamb_in1k`](https://huggingface.co/timm/mobilenetv2_050.lamb_in1k) for [kerasformers](https://github.com/IMvision12/KerasFormers). One implementation runs unmodified on **TensorFlow / Torch / JAX**.

This is an **image-classification / backbone** checkpoint (`MobileNetV2ImageClassify` / `MobileNetV2Model`).

## ✨ Quick start

```python
import os
os.environ["KERAS_BACKEND"] = "torch"  # or "jax" / "tensorflow"

from PIL import Image
import numpy as np
from kerasformers.models.mobilenetv2 import MobileNetV2ImageClassify, MobileNetV2Model

model = MobileNetV2ImageClassify.from_weights("kerasformers/mobilenetv2_050_lamb_in1k")
backbone = MobileNetV2Model.from_weights(
    "kerasformers/mobilenetv2_050_lamb_in1k", as_backbone=True
)

image = Image.open("your_image.jpg").convert("RGB")
image = image.resize((224, 224))
x = np.asarray(image, dtype="float32")[None]  # (1, H, W, 3)
print(model(x).shape)  # (1, num_classes)
feats = backbone(x)
print(len(feats), [tuple(f.shape) for f in feats])
```

Load any MobileNetV2 variant the same way with `from_weights("kerasformers/<variant>")`:

| Variant | Hub |
|---|---|
| `mobilenetv2_050_lamb_in1k` | [`kerasformers/mobilenetv2_050_lamb_in1k`](https://huggingface.co/kerasformers/mobilenetv2_050_lamb_in1k) |
| `mobilenetv2_100_ra_in1k` | [`kerasformers/mobilenetv2_100_ra_in1k`](https://huggingface.co/kerasformers/mobilenetv2_100_ra_in1k) |
| `mobilenetv2_110d_ra_in1k` | [`kerasformers/mobilenetv2_110d_ra_in1k`](https://huggingface.co/kerasformers/mobilenetv2_110d_ra_in1k) |
| `mobilenetv2_120d_ra_in1k` | [`kerasformers/mobilenetv2_120d_ra_in1k`](https://huggingface.co/kerasformers/mobilenetv2_120d_ra_in1k) |
| `mobilenetv2_140_ra_in1k` | [`kerasformers/mobilenetv2_140_ra_in1k`](https://huggingface.co/kerasformers/mobilenetv2_140_ra_in1k) |

## Tips

- Set `KERAS_BACKEND` **before** importing Keras / kerasformers.
- `MobileNetV2ImageClassify` returns class logits; `MobileNetV2Model` returns features (`as_backbone=True` for multi-scale stages).
- See [docs](https://imvision12.github.io/KerasFormers/classification_backbones/) and [Loading Weights](https://imvision12.github.io/KerasFormers/loading_weights/).
- Upstream / timm checkpoints: `MobileNetV2ImageClassify.from_weights("hf:timm/mobilenetv2_050.lamb_in1k")`.

## Special Thanks

A huge thank you to the MobileNetV2 authors and the timm / Hub communities for creating and releasing these models.

License: see YAML `license` (usually matches the upstream checkpoint).