RegNetY-008 Image Classification Model
| Property | Value |
|---|---|
| Parameter Count | 6.26M |
| Model Type | Image Classification |
| License | MIT |
| Paper | Designing Network Design Spaces |
| Framework | PyTorch (timm) |
What is regnety_008.pycls_in1k?
RegNetY-008 is a lightweight convolutional neural network designed for image classification tasks, developed as part of the RegNet family of models. This particular implementation achieves 76.296% top-1 accuracy on ImageNet-1k, making it suitable for resource-constrained applications while maintaining reasonable performance.
Implementation Details
The model is implemented in the timm library and features several architectural enhancements including stochastic depth, gradient checkpointing, and layer-wise learning rate decay. It operates on 224x224 pixel images and requires only 0.81 GMACs (billion multiply-accumulate operations) for inference.
- Efficient architecture with 6.26M parameters
- Supports configurable activation and normalization layers
- Features pre-activation bottleneck blocks
- Optimized for both accuracy and computational efficiency
Core Capabilities
- Image classification on ImageNet-1k dataset
- Feature extraction for downstream tasks
- Support for both classification and embedding generation
- Configurable output stride through dilation
Frequently Asked Questions
Q: What makes this model unique?
RegNetY-008 represents an optimal balance between model size and performance, derived from systematic network design space exploration. It's particularly notable for its efficiency, requiring minimal computational resources while maintaining reasonable accuracy.
Q: What are the recommended use cases?
This model is ideal for applications requiring lightweight image classification, such as mobile devices or edge computing scenarios. It's also suitable as a feature extractor for transfer learning tasks where computational resources are limited.
