Acly / krita-ai-tools

Krita plugin which adds selection tools to mask objects with a single click, or by drawing a bounding box.
GNU General Public License v3.0
373 stars 17 forks source link

A regular plugin version #18

Open vivi-gomez opened 2 months ago

vivi-gomez commented 2 months ago

Is there any plan to become a regular plugin installation? This is a great project and very useful. However, its installation it is complicated for a regular Krita app image with other plugin.

Acly commented 2 months ago

What is a "regular" plugin? If you mean Python plugin, that is limited and does not allow adding new tools (for example).

MurzNN commented 2 months ago

Yeah, a simplified installation would be great. If no tools can be added, maybe make just a separate docker with the plugin functionality? Here is an example of an AI plugin that works as a regular plugin: https://github.com/Acly/krita-ai-diffusion/

Acly commented 2 months ago

It's not a good option, apart from bad usability, the tools also inherit a lot of functionality (including all the tool options) from the other selection tools in krita. A python plugin would have to reimplement all that by hand.

Technically its possible to use the python plugin to load the SO from anywhere, but I tried this before and it didn't work (probably python plugin is loaded too late).

vivi-gomez commented 2 months ago

What is a "regular" plugin? If you mean Python plugin, that is limited and does not allow adding new tools (for example).

Well, I am just a user, in love with Krita and I amazed by krita-ai-tools. However, I am not able to make the installation coexist by opening the app file, installing and packaging it again to also have krita-ai-diffusion installed, as @MurzNN mentions. I don't know how it should be reprogrammed, but it's actually difficult to understand the installation.

Moreover, I do not why but I can use it only with the CPU, the GPU give me an error (can't say right now exactly what is problem)

Acly commented 2 months ago

I don't see how this is related to using krita-ai-diffusion? It lives in a completely different place (usually ~/.local/share/krita) and will work with the original AppImage or the modified version.

I know installation on Linux isn't ideal, but realistically the way to improve it is to get the feature into Krita.

Moreover, I do not why but I can use it only with the CPU, the GPU give me an error (can't say right now exactly what is problem)

Using GPU on Linux requires installing a bunch of NVIDIA stuff, none of which is available as packages. It's not worth the effort.

vivi-gomez commented 2 months ago

I don't see how this is related to using krita-ai-diffusion? It lives in a completely different place (usually ~/.local/share/krita) and will work with the original AppImage or the modified version.

I know installation on Linux isn't ideal, but realistically the way to improve it is to get the feature into Krita.

Moreover, I do not why but I can use it only with the CPU, the GPU give me an error (can't say right now exactly what is problem)

Using GPU on Linux requires installing a bunch of NVIDIA stuff, none of which is available as packages. It's not worth the effort.

That is the thing. I could not manage to have both things together. I am running two instances of Krita in order to use both things. My GPU is working fine with all other software, I thought I did a wrong installation of krita-ai-tools since I only able to use the cpu. Thank you for your support

moonwhaler commented 2 months ago

I, too, would love this, but I understand the sentiment and the effort involved.

Currently, I am (Linux) distro hopping a bit, which came to a stop with Arch. Not using Krita Flatpaks solved many issues I had with using this project. In fact, it was almost as easy as in Windows.

What would really be a shame, if this project would be abandoned. It's a real help!

yen-shi-lun commented 3 weeks ago

Thank you very much for developing this plugin, 5.2.3 works fine

xiaojianan123 commented 1 week ago

非常感谢您开发这个插件,5.2.3 运行良好

Why Can't I run