Models for megengine
The Correspondence:
| path | name in code | paper title |
|---|---|---|
| dncnn | dncnn_25 |
Beyond a Gaussian Denoiser: Residual Learning of Deep CNN for Image Denoising |
| n2n | n2n_gauss25, n2n_gauss5to50, n2n_poisson30, n2n_poisson5to50 |
Neighbor2Neighbor: Self-Supervised Denoising from Single Noisy Images |
| nts | attention_net |
Learning to Navigate for Fine-grained Classification |
| esrgan | rrdb_psnr, rrdb_esrgan |
ESRGAN: Enhanced Super-Resolution Generative Adversarial Networks |
| rcan | rcan_x2, rcan_x3, rcan_x4 |
Image Super-Resolution Using Very Deep Residual Channel Attention Networks |
| csd | csd_edsr_mix, csd_edsr_student |
Towards Compact Single Image Super-Resolution via Contrastive Self-distillation |
| rdn | rdn_x2, rdn_x3, rdn_x4 |
Residual Dense Network for Image Super-Resolution |
Pretrained pytorch models:
- Link: https://pan.baidu.com/s/1-pw6tUwmU0Q0uFbU-K2p1A code: y1md
Thanks to the following authors for generously making their code available.
- DnCNN : https://github.com/cszn/DnCNN
- N2N : https://github.com/TaoHuang2018/Neighbor2Neighbor
- NTS-Net : https://github.com/yangze0930/NTS-Net
- ESRGAN : https://github.com/xinntao/ESRGAN
- RCAN : https://github.com/yulunzhang/RCAN
- CSD : https://github.com/Booooooooooo/CSD
- RDN : https://github.com/yulunzhang/RDN https://github.com/open-mmlab/mmediting