Skip to content

Latest commit

 

History

History
150 lines (108 loc) · 8.52 KB

ModelZoo.md

File metadata and controls

150 lines (108 loc) · 8.52 KB
Raw

Model Zoo and Baselines

English | 简体中文

Download: ⏬ Google Drive: Pretrained Models | Reproduced Experiments ⏬ 百度网盘: 预训练模型 | 复现实验

📈 Training curves in wandb

We provide:

  1. Official models converted directly from official released models
  2. Reproduced models with BasicSR. Pre-trained models and log examples are provided

You can put the downloaded models in the experiments/pretrained_models folder.

[Download official pre-trained models] (Google Drive, 百度网盘)

You can use the script to download pre-trained models from Google Drive.

python scripts/download_pretrained_models.py ESRGAN
# method can be ESRGAN, EDVR, StyleGAN, EDSR, DUF, DFDNet, dlib

[Download reproduced models and logs] (Google Drive, 百度网盘)

In addition, we upload the training process and curves in wandb.

Training curves in wandb

Contents

  1. Image Super-Resolution
    1. Image SR Official Models
    2. Image SR Reproduced Models
  2. Video Super-Resolution

Image Super-Resolution

When evaluation:

  • We crop scale border pixels in each border
  • Evaluated on RGB channels

Image SR Official Models

Exp Name Set5 (PSNR/SSIM) Set14 (PSNR/SSIM) DIV2K100 (PSNR/SSIM)
EDSR_Mx2_f64b16_DIV2K_official-3ba7b086 35.7768 / 0.9442 31.4966 / 0.8939 34.6291 / 0.9373
EDSR_Mx3_f64b16_DIV2K_official-6908f88a 32.3597 / 0.903 28.3932 / 0.8096 30.9438 / 0.8737
EDSR_Mx4_f64b16_DIV2K_official-0c287733 30.1821 / 0.8641 26.7528 / 0.7432 28.9679 / 0.8183
EDSR_Lx2_f256b32_DIV2K_official-be38e77d 35.9979 / 0.9454 31.8583 / 0.8971 35.0495 / 0.9407
EDSR_Lx3_f256b32_DIV2K_official-3660f70d 32.643 / 0.906 28.644 / 0.8152 31.28 / 0.8798
EDSR_Lx4_f256b32_DIV2K_official-76ee1c8f 30.5499 / 0.8701 27.0011 / 0.7509 29.277 / 0.8266

Image SR Reproduced Models

Experiment name conventions are in Config.md.

Exp Name Set5 (PSNR/SSIM) Set14 (PSNR/SSIM) DIV2K100 (PSNR/SSIM)
001_MSRResNet_x4_f64b16_DIV2K_1000k_B16G1_wandb 30.2468 / 0.8651 26.7817 / 0.7451 28.9967 / 0.8195
002_MSRResNet_x2_f64b16_DIV2K_1000k_B16G1_001pretrain_wandb 35.7483 / 0.9442 31.5403 / 0.8937 34.6699 / 0.9377
003_MSRResNet_x3_f64b16_DIV2K_1000k_B16G1_001pretrain_wandb 32.4038 / 0.9032 28.4418 / 0.8106 30.9726 / 0.8743
004_MSRGAN_x4_f64b16_DIV2K_400k_B16G1_wandb 28.0158 / 0.8087 24.7474 / 0.6623 26.6504 / 0.7462
201_EDSR_Mx2_f64b16_DIV2K_300k_B16G1_wandb 35.7395 / 0.944 31.4348 / 0.8934 34.5798 / 0.937
202_EDSR_Mx3_f64b16_DIV2K_300k_B16G1_201pretrain_wandb 32.315 / 0.9026 28.3866 / 0.8088 30.9095 / 0.8731
203_EDSR_Mx4_f64b16_DIV2K_300k_B16G1_201pretrain_wandb 30.1726 / 0.8641 26.721 / 0.743 28.9506 / 0.818
204_EDSR_Lx2_f256b32_DIV2K_300k_B16G1_wandb 35.9792 / 0.9453 31.7284 / 0.8959 34.9544 / 0.9399
205_EDSR_Lx3_f256b32_DIV2K_300k_B16G1_204pretrain_wandb 32.6467 / 0.9057 28.6859 / 0.8152 31.2664 / 0.8793
206_EDSR_Lx4_f256b32_DIV2K_300k_B16G1_204pretrain_wandb 30.4718 / 0.8695 26.9616 / 0.7502 29.2621 / 0.8265

Video Super-Resolution

Evaluation

In the evaluation, we include all the input frames and do not crop any border pixels unless otherwise stated.
We do not use the self-ensemble (flip testing) strategy and any other post-processing methods.

EDVR

Name convention
EDVR_(training dataset)_(track name)_(model complexity)

  • track name. There are four tracks in the NTIRE 2019 Challenges on Video Restoration and Enhancement:
    • SR: super-resolution with a fixed downsampling kernel (MATLAB bicubic downsampling kernel is frequently used). Most of the previous video SR methods focus on this setting.
    • SRblur: the inputs are also degraded with motion blur.
    • deblur: standard deblurring (motion blur).
    • deblurcomp: motion blur + video compression artifacts.
  • model complexity
    • L (Large): # of channels = 128, # of back residual blocks = 40. This setting is used in our competition submission.
    • M (Moderate): # of channels = 64, # of back residual blocks = 10.
Model name [Test Set] PSNR/SSIM
EDVR_Vimeo90K_SR_L [Vid4] (Y1) 27.35/0.8264 [↓Results]
(RGB) 25.83/0.8077
EDVR_REDS_SR_M [REDS] (RGB) 30.53/0.8699 [↓Results]
EDVR_REDS_SR_L [REDS] (RGB) 31.09/0.8800 [↓Results]
EDVR_REDS_SRblur_L [REDS] (RGB) 28.88/0.8361 [↓Results]
EDVR_REDS_deblur_L [REDS] (RGB) 34.80/0.9487 [↓Results]
EDVR_REDS_deblurcomp_L [REDS] (RGB) 30.24/0.8567 [↓Results]

1 Y or RGB denotes the evaluation on Y (luminance) or RGB channels.

Stage 2 models for the NTIRE19 Competition

Model name [Test Set] PSNR/SSIM
EDVR_REDS_SR_Stage2 [REDS] (RGB) / [↓Results]
EDVR_REDS_SRblur_Stage2 [REDS] (RGB) / [↓Results]
EDVR_REDS_deblur_Stage2 [REDS] (RGB) / [↓Results]
EDVR_REDS_deblurcomp_Stage2 [REDS] (RGB) / [↓Results]

DUF

The models are converted from the officially released models.

Model name [Test Set] PSNR/SSIM1 Official Results2
DUF_x4_52L_official3 [Vid4] (Y4) 27.33/0.8319 [↓Results]
(RGB) 25.80/0.8138		 (Y) 27.33/0.8318 [↓Results]
(RGB) 25.79/0.8136
DUF_x4_28L_official [Vid4]
DUF_x4_16L_official [Vid4]
DUF_x3_16L_official [Vid4]
DUF_x2_16L_official [Vid4]

1 We crop eight pixels near image boundary for DUF due to its severe boundary effects.
2 The official results are obtained by running the official codes and models.
3 Different from the official codes, where zero padding is used for border frames, we use new_info strategy.
4 Y or RGB denotes the evaluation on Y (luminance) or RGB channels.

TOF

The models are converted from the officially released models.

Model name [Test Set] PSNR/SSIM Official Results1
TOF_official2 [Vid4] (Y3) 25.86/0.7626 [↓Results]
(RGB) 24.38/0.7403		 (Y) 25.89/0.7651 [↓Results]
(RGB) 24.41/0.7428

1 The official results are obtained by running the official codes and models. Note that TOFlow does not provide a strategy for border frame recovery and we simply use a replicate strategy for border frames.
2 The converted model has slightly different results, due to different implementation. And we use new_info strategy for border frames.
3 Y or RGB denotes the evaluation on Y (luminance) or RGB channels.