Hello! Is your project still alive? It would've been really awesome to see the source code of what I played with :)
Thanks for you interest.
Yes, the code will be published soon (22th october 2018 at the latest).
We're almost done cleaning up the code and restructuring things.
For the open release, I'll try to manage training a much deeper model successfully. (The one which was online had only seven residual blocks).
Currently, I'm at a different place. I'll start the training as soon as I will have access to my graphics card again.
Yes, the code will be published soon (22th october 2018 at the latest).
We're almost done cleaning up the code and restructuring things.
For the open release, I'll try to manage training a much deeper model successfully. (The one which was online had only seven residual blocks).
Currently, I'm at a different place. I'll start the training as soon as I will have access to my graphics card again.
I'm happy to announce that the first official release of CrazyAra 0.2 is finally out. The new version was trained supervised only using a deeper novel architecture on the same dataset as before.
You can download it at and run it as an UCI-engine:
* github.com/QueensGambit/CrazyAra/releases
For more details about the project you can visit the wiki:
* github.com/QueensGambit/CrazyAra/wiki
You can download it at and run it as an UCI-engine:
* github.com/QueensGambit/CrazyAra/releases
For more details about the project you can visit the wiki:
* github.com/QueensGambit/CrazyAra/wiki
Let us know when it goes live! If you message me, I can also share when it's live again to the chess variants discord: dscrd.me/housechessvariants
This topic has been archived and can no longer be replied to.