DLP 26T2 NPPE-3 — Low-Light Super-Resolution
Lightweight SwinIR-style model trained for low-light image denoising and 4x super-resolution.
Model
- Architecture: SwinIRLite
- Input: RGB LR image
- Input resolution: 256 × 160
- Output resolution: 1024 × 640
- Upscaling factor: 4x
- Parameters: 1,516,995
Performance
Kaggle public leaderboard PSNR: 38.69453 dB
Files
model.pth— PyTorch model weightsconfig.json— model architecture and configuration
- Downloads last month
- -
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support