| Property | Value |
|---|---|
| Parameter Count | 0.5M |
| Model Type | Image Classification / Feature Backbone |
| Input Resolution | 160 x 160 |
| GMACs | 0.1 |
| Activations | 0.6M |
| Dataset | ImageNet-1k |
| Author | Ross Wightman |
What is test_resnet.r160_in1k?
test_resnet.r160_in1k is a compact ResNet model specifically designed for testing and sanity checks in computer vision applications. Developed by Ross Wightman as part of the PyTorch Image Models (timm) collection, this model represents a lightweight implementation trained on the ImageNet-1k dataset.
Implementation Details
The model features a streamlined architecture with just 0.5M parameters, making it extremely efficient while maintaining reasonable performance. It operates on 160x160 pixel inputs and requires only 0.1 GMACs for inference, with activation memory of 0.6M.
- Efficient parameter utilization with only 0.5M parameters
- Compact input resolution of 160x160 pixels
- Low computational overhead at 0.1 GMACs
- Achieves 41.578% top-1 accuracy on ImageNet
Core Capabilities
- Image classification with 1000 ImageNet classes
- Feature extraction for downstream tasks
- Embedding generation for transfer learning
- Support for both 160x160 and 192x192 input resolutions
Frequently Asked Questions
Q: What makes this model unique?
This model's primary strength lies in its minimal footprint while maintaining functionality for testing purposes. It's specifically designed for validation and debugging of machine learning pipelines rather than production use.
Q: What are the recommended use cases?
The model is best suited for testing environments, proof-of-concept implementations, and educational purposes where a lightweight ResNet implementation is needed. It's particularly useful for validating data processing pipelines and debugging deep learning workflows.