README.md

metadata

license: cc-by-4.0
pipeline_tag: image-to-image
tags:
  - pytorch
  - super-resolution

Link to Github Release

2xHFA2kAVCEDSR_M

Name: 2xHFA2kAVCEDSR_M
Author: Philip Hofmann
Release Date: 18.06.2023
License: CC BY 4.0
Network: EDSR M
Scale: 2
Purpose: Fast 2x anime upscaling model that handles AVC (h264) degradation
Iterations: 186,000
batch_size: 2-5
HR_size: 384
Epoch: 317 (require iter number per epoch: 1284)
Dataset: HFA2k_h264
Number of train images: 2568
OTF Training: No
Pretrained_Model_G: None

Description: A 2x EDSR M anime upscale model that handles AVC (h264) degradation. Applied h264 crf 20-28 degradation and bicubic, bilinear, box and lanczos downsampling on the HFA2k dataset with Kim's dataset destroyer.