I used PGX and MCTX to train AlphaZero on Kuhn Poker. It ran on a TPU v2-8(courtesy of the TPU Research Cloud Program) for ~3.5 days. Code can be found here.
- Downloads last month
- 0
Unable to determine this model's library. Check the
docs
.
I used PGX and MCTX to train AlphaZero on Kuhn Poker. It ran on a TPU v2-8(courtesy of the TPU Research Cloud Program) for ~3.5 days. Code can be found here.