3 ms·
I meant trivial in the sense it's a solved problem, I'm sure it still costs a non-negligible amount of money to train it. See for example the chess transformer
by pu_pe 7mo ago
I meant trivial in the sense it's a solved problem, I'm sure it still costs a non-negligible amount of money to train it. See for example the chess transformer built by DeepMind a couple of years ago which I referred to in a sibling comment [1].
[1] https://arxiv.org/abs/2402.04494 https://arxiv.org/abs/2402.04494
- Otterly99 7mo agoThank you for the link. I admit, my knowledge of reinforcement learning is a bit outdated so it seemed to me that it was unattainable for a non-specialized model to train efficiently on something like chess, which has a huge state space.