Hugging Face's logo

Spaces:
Sumit7864
/
Image-Enhancer
Running

App Files Community
1
Image-Enhancer / docs /anime_video_model.md
Rakesh Chavhan
inital commit
8875fed
preview code
|
raw
history blame
No virus
5.72 kB

Anime Video Models

:white_check_mark: We add small models that are optimized for anime videos :-)
More comparisons can be found in anime_comparisons.md

Models Scale Description
realesr-animevideov3 X4 1 Anime video model with XS size

Note:
1 This model can also be used for X1, X2, X3.

The following are some demos (best view in the full screen mode).

https://user-images.githubusercontent.com/17445847/145706977-98bc64a4-af27-481c-8abe-c475e15db7ff.MP4

https://user-images.githubusercontent.com/17445847/145707055-6a4b79cb-3d9d-477f-8610-c6be43797133.MP4

https://user-images.githubusercontent.com/17445847/145783523-f4553729-9f03-44a8-a7cc-782aadf67b50.MP4

How to Use

PyTorch Inference 

# download model
wget https://github.com/xinntao/Real-ESRGAN/releases/download/v0.2.5.0/realesr-animevideov3.pth -P weights
# single gpu and single process inference
CUDA_VISIBLE_DEVICES=0 python inference_realesrgan_video.py -i inputs/video/onepiece_demo.mp4 -n realesr-animevideov3 -s 2 --suffix outx2
# single gpu and multi process inference (you can use multi-processing to improve GPU utilization)
CUDA_VISIBLE_DEVICES=0 python inference_realesrgan_video.py -i inputs/video/onepiece_demo.mp4 -n realesr-animevideov3 -s 2 --suffix outx2 --num_process_per_gpu 2
# multi gpu and multi process inference
CUDA_VISIBLE_DEVICES=0,1,2,3 python inference_realesrgan_video.py -i inputs/video/onepiece_demo.mp4 -n realesr-animevideov3 -s 2 --suffix outx2 --num_process_per_gpu 2

Usage:
--num_process_per_gpu    The total number of process is num_gpu * num_process_per_gpu. The bottleneck of
                         the program lies on the IO, so the GPUs are usually not fully utilized. To alleviate
                         this issue, you can use multi-processing by setting this parameter. As long as it
                         does not exceed the CUDA memory
--extract_frame_first    If you encounter ffmpeg error when using multi-processing, you can turn this option on.

NCNN Executable File

Step 1: Use ffmpeg to extract frames from video 

ffmpeg -i onepiece_demo.mp4 -qscale:v 1 -qmin 1 -qmax 1 -vsync 0 tmp_frames/frame%08d.png
  • Remember to create the folder tmp_frames ahead

Step 2: Inference with Real-ESRGAN executable file

  1. Download the latest portable Windows / Linux / MacOS executable files for Intel/AMD/Nvidia GPU

  2. Taking the Windows as example, run:

    ./realesrgan-ncnn-vulkan.exe -i tmp_frames -o out_frames -n realesr-animevideov3 -s 2 -f jpg
    • Remember to create the folder out_frames ahead

Step 3: Merge the enhanced frames back into a video

  1. First obtain fps from input videos by

    ffmpeg -i onepiece_demo.mp4

    Usage:
-i                   input video path

    You will get the output similar to the following screenshot.

  2. Merge frames

    ffmpeg -r 23.98 -i out_frames/frame%08d.jpg -c:v libx264 -r 23.98 -pix_fmt yuv420p output.mp4

    Usage:
-i                   input video path
-c:v                 video encoder (usually we use libx264)
-r                   fps, remember to modify it to meet your needs
-pix_fmt             pixel format in video

    If you also want to copy audio from the input videos, run:

     ffmpeg -r 23.98 -i out_frames/frame%08d.jpg -i onepiece_demo.mp4 -map 0:v:0 -map 1:a:0 -c:a copy -c:v libx264 -r 23.98 -pix_fmt yuv420p output_w_audio.mp4

    Usage:
-i                   input video path, here we use two input streams
-c:v                 video encoder (usually we use libx264)
-r                   fps, remember to modify it to meet your needs
-pix_fmt             pixel format in video

More Demos

More comparisons

https://user-images.githubusercontent.com/17445847/145707458-04a5e9b9-2edd-4d1f-b400-380a72e5f5e6.MP4