loading G1 + policy + physics (WASM)…
nanoG1
Training duration:  58.9s
Samples:  75M @ 1.28M SPS
Compute:  1× RTX PRO 6000
By Julian Saks
arrows walk/turn · R reset