Update README.md
Browse files
README.md
CHANGED
|
@@ -13,3 +13,8 @@ This repository contains the trained weights and architecture for a Hybrid Atten
|
|
| 13 |
* `hat_phase2.pth`: The final fine-tuned PyTorch model weights.
|
| 14 |
* `hat_arch.py`: The architecture definition used to instantiate the model.
|
| 15 |
* `config.json`: The exact hyperparameters required to construct the architecture.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 13 |
* `hat_phase2.pth`: The final fine-tuned PyTorch model weights.
|
| 14 |
* `hat_arch.py`: The architecture definition used to instantiate the model.
|
| 15 |
* `config.json`: The exact hyperparameters required to construct the architecture.
|
| 16 |
+
|
| 17 |
+
## Evaluation
|
| 18 |
+
**Note on Reproducibility:** The reported PSNR of 39.57 is achieved using **8-fold Test-Time Augmentation (TTA)** combined with overlapping patch inference (tile size 128) to prevent memory constraints.
|
| 19 |
+
|
| 20 |
+
Running a single raw forward pass without TTA will result in a slightly lower PSNR. Please refer to the provided inference script/notebook for the exact TTA implementation.
|