WideResNet50: Image Classification
WideResNet50 is an enhanced residual network that boosts performance by increasing network width (channel count) rather than depth. It employs wider residual blocks (e.g., width factor of 2), expanding feature dimensions while reducing layers, balancing computational efficiency and representational power. Retaining residual skip connections to mitigate vanishing gradients, it uses batch normalization for faster convergence. Compared to ResNet-50, WideResNet50 achieves higher accuracy on datasets like ImageNet with controlled parameter growth, suitable for image classification and object detection. Its design prioritizes "width over depth," ideal for resource-constrained yet accuracy-demanding applications.
Source model
- Input shape: 640x640
- Number of parameters: 4.44M
- Model size: 17.91 MB
- Output shape: 1x8400x85
The source model can be found here
Performance Reference
Please search model by model name in Model Farm
Inference & Model Conversion
Please search model by model name in Model Farm
License
Source Model: BSD-3-CLAUSE
Deployable Model: APLUX-MODEL-FARM-LICENSE