simsim314

OK I see everything's there, except that it doesn't work well. I think from practical standpoint, we should reach a place where alpha-zero is not giving away it's queen or...

I think we should be realistic about our access to good hardware. Google had used 5000 TPUs to generate self play, I think we can safely assume we will not...

Is it possible to visualize the hyperbolic space in more user friendly manner? Hyperbolic CA could be interesting - if we would have nicer visualization for it. I mean in...

Hmm how about for starters "size regularization" stretching the radius so that every new layer of N-gons will be optimized to be the size of the rest of the grid....

I'm not sure what projection it is. I want something more or less simple: O(1) approachable adjacency graph. Preferably calculable by function and not stored in matrix of adjacencies -...

I just need very effective visualization of the hyperbolic plane and very computationally cheap neighbor function all represented in 2d grid locally at least not so much stretched. So that...

Here is my explanation using golly: https://www.twitch.tv/videos/392238658

As far as I understand those two are hardcoded keywords that make the alpaca a chat from llama which is a text completion model. [code of "### Instruction"](https://github.com/antimatter15/alpaca.cpp/blob/master/chat.cpp#L860) [code of...

It's possible to avoid token associated with "##" Another option is during reinforcement training to give negative reward for such outputs.