๐ ReXNet-1.5x model
This model is pretrained on ImageNette. The ReXNet architecture was introduced in this paper. It is designed for image classification tasks using PyTorch and ONNX.
๐ Quick Start
The ReXNet-1.5x model is pretrained on ImageNette. It aims to address the representational bottleneck in convolutional neural networks.
โจ Features
- Pretrained on the ImageNette dataset.
- The core idea involves adding a customized Squeeze - Excitation layer in residual blocks to prevent channel redundancy.
๐ฆ Installation
Prerequisites
Python 3.6 (or higher) and pip/conda are required to install Holocron.
Latest stable release
You can install the last stable release of the package using pypi as follows:
pip install pylocron
or using conda:
conda install -c frgfm pylocron
Developer mode
Alternatively, if you wish to use the latest features of the project that haven't made their way to a release yet, you can install the package from source (install Git first):
git clone https://github.com/frgfm/Holocron.git
pip install -e Holocron/.
๐ป Usage Examples
Basic Usage
from PIL import Image
from torchvision.transforms import Compose, ConvertImageDtype, Normalize, PILToTensor, Resize
from torchvision.transforms.functional import InterpolationMode
from holocron.models import model_from_hf_hub
model = model_from_hf_hub("frgfm/rexnet1_5x").eval()
img = Image.open(path_to_an_image).convert("RGB")
config = model.default_cfg
transform = Compose([
Resize(config['input_shape'][1:], interpolation=InterpolationMode.BILINEAR),
PILToTensor(),
ConvertImageDtype(torch.float32),
Normalize(config['mean'], config['std'])
])
input_tensor = transform(img).unsqueeze(0)
with torch.inference_mode():
output = model(input_tensor)
probs = output.squeeze(0).softmax(dim=0)
๐ Documentation
The core idea of the author is to add a customized Squeeze - Excitation layer in the residual blocks that will prevent channel redundancy.
๐ License
This project is licensed under the Apache - 2.0 license.
๐ Citation
Original paper
@article{DBLP:journals/corr/abs-2007-00992,
author = {Dongyoon Han and
Sangdoo Yun and
Byeongho Heo and
Young Joon Yoo},
title = {ReXNet: Diminishing Representational Bottleneck on Convolutional Neural
Network},
journal = {CoRR},
volume = {abs/2007.00992},
year = {2020},
url = {https://arxiv.org/abs/2007.00992},
eprinttype = {arXiv},
eprint = {2007.00992},
timestamp = {Mon, 06 Jul 2020 15:26:01 +0200},
biburl = {https://dblp.org/rec/journals/corr/abs-2007-00992.bib},
bibsource = {dblp computer science bibliography, https://dblp.org}
}
Source of this implementation
@software{Fernandez_Holocron_2020,
author = {Fernandez, Franรงois - Guillaume},
month = {5},
title = {{Holocron}},
url = {https://github.com/frgfm/Holocron},
year = {2020}
}
Property |
Details |
Model Type |
Image Classification Model |
Training Data |
frgfm/imagenette |