DeOldify/VideoColorizerColab.ipynb

#◢ DeOldify - Not just for photos!

##Use this Colab notebook to colorize black & white videos in three simple steps.

1. Specify media URL - YouTube, Twitter, Imgur, etc.
2. Run DeOldify to extract single images from your video or gif. Behind the scenes, the code does the following:
   • Extracts single images from the specified media file.
   • Processes the images with DeOldify.
   • Rebuilds the video from colorized images.
3. Download the video to your device to view!

FYI: This notebook is intended as a tool to colorize gifs and short videos, if you are trying to convert longer video you may hit the limit on processing space. Running the Jupyter notebook on your own machine is recommended (and faster) for larger video sizes.

---

####Credits:

Use of this tool is thanks to,

@citnaj for creating DeOldify.

@tradica for initial video and CoLab work.

Matt Robinson for his notebook which helped make DeOldify approachable.

Dana Kelley for doing things, breaking stuff & having an opinion on everything.

---

#◢ Verify Correct Runtime Settings

IMPORTANT

In the "Runtime" menu for the notebook window, select "Change runtime type." Ensure that the following are selected:

• Runtime Type = Python 3
• Hardware Accelerator = GPU

In [ ]:

from os import path
import torch
print(torch.__version__)
print(torch.cuda.is_available())

#◢ Git clone and install DeOldify

In [ ]:

!git clone -b FastAIv1 --single-branch https://github.com/jantic/DeOldify.git DeOldify
#!git clone https://github.com/jantic/DeOldify.git DeOldify

In [ ]:

cd DeOldify

#◢ Setup

In [ ]:

!pip install PyDrive
!pip install ffmpeg-python
!pip install youtube-dl
!pip install tensorboardX

In [ ]:

import fastai
from fasterai.visualize import *
from pathlib import Path
from google.colab import drive
from google.colab import files
torch.backends.cudnn.benchmark=True

In [ ]:

!mkdir 'models'
!wget https://www.dropbox.com/s/ztgygpaz1z3jkjg/ColorizeImagesStable_gen.pth?dl=0 -O ./models/ColorizeImagesStable_gen.pth

#◢ Render Factor

The default value of 21 has been carefully chosen and should work for most scenarios. This determines resolution at which video is rendered. Lower resolution will render faster, and colors also tend to look more vibrant. Older and lower quality film in particular will generally benefit by lowering the render factor. Higher render factors are often better for higher quality videos and inconsistencies (flashy render) will generally be reduced, but the colors may get slightly washed out.

In [ ]:

render_factor = 21  #@param {type: "slider", min: 5, max: 45}

#◢ Specify URL

YouTube, Imgur, Twitter, Reddit ... files of type .gif, .gifv and .mp4 work. NOTE: If you want to use your own source material, upload it first to a site like Imgur.

In [ ]:

source_url = '' #@param {type:"string"}

#◢ Additional Parameters

It's not necessary to change the following, just run them as-is.

In [ ]:

file_name = 'video.mp4'
source_dir = './video/source/'
source_path = source_dir + file_name
dest_dir = './video/result/'
dest_path = dest_dir + file_name

In [ ]:

!mkdir file_dir

#◢ Run DeOldify

In [ ]:

colorizer = get_video_colorizer(render_factor=render_factor)

In [ ]:

if source_url is not None and source_url !='':
    colorizer.colorize_from_url(source_url, file_name)
else:
    print('Provide a source url and try again.')

#◢ Download

• In the menu to the left, click Files
• If you don't see the 'DeOldify' folder, click "Refresh"
• By default, rendered video will be in /DeOldify/video/result/

---

#⚙ Recommended video and gif sources

• /r/Nickelodeons/
• https://twitter.com/silentmoviegifs