Regnet Y 320

🚀 RegNet

RegNet is a model trained on ImageNet-1K, designed to address image classification tasks and offering high - performance solutions for the field of computer vision.

🚀 Quick Start

RegNet is a powerful model for image classification. You can use the raw model directly for this task, or explore fine - tuned versions on the model hub.

✨ Features

• Search Space Design: The authors design search spaces for Neural Architecture Search (NAS), iteratively reducing the search space based on the best - performing models sampled from the current space.
• Trained on ImageNet - 1K: It has been trained on the large - scale ImageNet - 1K dataset, which enhances its generalization ability in image classification tasks.

📚 Documentation

Model description

The authors design search spaces to perform Neural Architecture Search (NAS). They first start from a high dimensional search space and iteratively reduce the search space by empirically applying constraints based on the best - performing models sampled by the current search space.

Intended uses & limitations

You can use the raw model for image classification. See the model hub to look for fine - tuned versions on a task that interests you.

💻 Usage Examples

Basic Usage

>>> from transformers import AutoFeatureExtractor, RegNetForImageClassification
>>> import torch
>>> from datasets import load_dataset

>>> dataset = load_dataset("huggingface/cats-image")
>>> image = dataset["test"]["image"][0]

>>> feature_extractor = AutoFeatureExtractor.from_pretrained("zuppif/regnet-y-040")
>>> model = RegNetForImageClassification.from_pretrained("zuppif/regnet-y-040")

>>> inputs = feature_extractor(image, return_tensors="pt")

>>> with torch.no_grad():
...     logits = model(**inputs).logits

>>> # model predicts one of the 1000 ImageNet classes
>>> predicted_label = logits.argmax(-1).item()
>>> print(model.config.id2label[predicted_label])
'tabby, tabby cat'

Advanced Usage

For more code examples, we refer to the documentation.

📄 License

This project is licensed under the Apache - 2.0 license.