vicioussquid

Now that I have built this tool I'm going to use it to create a game

My USP is specifically targeting Snapdragon 8CX (a mobile CPU) and OpenGL 3.3 with Adreno graphics

Nope, all done real-time using Frustum culling.

`Baking` of lightmaps is supported but not currently necessary

Other than the 14 examples included, not yet; this is brand new and I am reaching out to the community to create some stuff with it, extend it with angled brushes, new entity types, etc

Thanks so much for your interest.

https://github.com/ViciousSquid/Dosidicus

This update introduces optional AI accelerator support via ONNX Runtime while keeping NumPy fallback.

I am looking for testers, thank you for your interest x

I posted an earlier version of my digital pet with a neural network about 10 months ago — since then I’ve added optional ONNX Runtime acceleration. No Pytorch.

Wow! This has blown up overnight! 150 stars on Github! Thanks everyone!

Have you seen any emergent behaviors from the squids that you didn't expect?

I have a squid that has become unnecessarily obsessed with playing with poop - throwing it around

Also another that likes to hoard decoration items by pushing them all together into a big pile

What if a Tamagotchi had a neural network and could learn stuff?

A digital pet squid that also teaches how neural networks and hebbian learning work. Behaviours are driven by the neural network according to his needs:

https://github.com/ViciousSquid/Dosidicus

I spent AGES on this and would love feedback. I think it's just the right balance of educational and fun. I did all the graphics myself and am currently working on multiplayer - squids will be able top enter other tanks and steal things, bring them home