Batch mode improvements


#1

The desktop app has been very useful for me. Thank you for your great work on it.

I have a suggestion: Batch mode would be my most used feature if it had a few tweaks. Namely:

  • Allow multiple input images
  • Allow selection of styles instead of all styles
  • Output each of the input images’ styles in its own folder.

For example:
Input: [input1.jpg, input2.jpg]
Styles: [Abstract 1, Colorful]
Options: [original_colors: true]
Output: [input1/abstract_original.jpg, input1/colorful_original.jpg, input2/abstract_original.jpg, input2/colorful_original.jpg]

The more granularity available for batch mode, the better.

Thanks for listening!


#2

Thank you for your feedback. We will discuss this features!


#3

It’s actually possible to batch process but it’s expensive

I already know how to do this but it involves using the Web APIs

I got Desktop today but it ended up telling me it was gonna take three weeks or so to process a 2:30 video I have for funding for a kickstarter

You can get 15,000 images converted per month if you obey the rules for $38 (more if you pay more)

I’m using Desktop to test this out (or thought I was)

Why ain’t there a command-line version?


#4

If you want fast processing in high quality you have to use the GPU Version and you need a Nvidia graphics card. That’s why turning images into art is not a trivial task. We set command line on our feature list. So far no one request it.


#5

I’ve got CUDA on my Windows10 Laptop DeepArtEffects doesn’t recognise it

cuda

As is common with Laptops mine defaults to some Intel GPU but switches to my NVidia GPU when it’s needed.

The result is DeepArtEffects doesn’t see Cuda so the GPU version is not functional

As for command line arguments - I tried starting it from the command prompt with a -v as this usually shows the version, the result was simply the same as clicking the icon on the desktop.

There is no documentation whatsoever similar to mobile apps as they rarely have any decent help)