Show HN: nanoAlphaZero – Train a grandmaster-level chess model in 24h with TPUsgithub.com·3 pts·tdoubleu·0