SESR-M5: Super Resolution
SESR-M5 is an efficient image super-resolution model designed for real-time quality enhancement via lightweight architecture. Built on deep convolutional networks, it optimizes multi-scale feature fusion and computational efficiency, supporting 2x to 4x upscaling for mobile or edge device deployment. It balances detail reconstruction and denoising under low-resource constraints, targeting artifacts, blur, and low-light degradation in applications like real-time video enhancement, historical media restoration, and mobile photography. Challenges include model compression for portability, cross-device compatibility, and stability in dynamic scenes.
Source model
- Input shape: 1x3x128x128
- Number of parameters: 0.32M
- Model size: 1.32M
- Output shape: 1x3x512x512
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: AIMET-MODEL-ZOO
Deployable Model: AIMET-MODEL-ZOO