Downloads · 30 days
6
21% of all-time downloads
Abhishek46777/dlp26t2-nppe3-exp4-rcan
dlp26t2-nppe3-exp4-rcan is a machine learning model from Abhishek46777. Use it for the machine learning task on the model card, and read the license before you ship it in a product.
This repository contains the trained model checkpoint used for the DLP26T2 NPPE-3 competition.
Downloads · 30 days
6
21% of all-time downloads
All-time downloads
28
Public
Repo size
28.4 MB
Likes
1
Public
Click a slice to open those files.
.pth28.4 MB · 100%
From the Hugging Face model README
This repository contains the trained model checkpoint used for the DLP26T2 NPPE-3 competition.
Given a low-resolution, noisy and low-light RGB image, reconstruct a clean high-resolution image with 4× spatial upscaling.
The model is a lightweight RCAN-inspired residual network with:
The final reconstruction is:
Prediction = Bicubic(LR) + Learned Residual
The model was trained using:
The uploaded checkpoint corresponds to the best validation PSNR obtained during training.
For the final competition predictions, four-way Test-Time Augmentation (TTA) was applied:
The transformed predictions were mapped back to the original orientation and averaged.
Validation PSNR:
The TTA procedure is applied during inference and is not part of the model checkpoint itself.
best_model_exp3.pth contains the PyTorch checkpoint with the model state dictionary and training metadata.