Evaluation sharing
Explore frozen-policy evaluations, recorded episodes, and the exact training checkpoints behind them.
| Name | Algorithm | Seed | Progress unit |
|---|---|---|---|
| 20M timesteps | ippo_rnn | 901 | steps |
| 213updates | ippo_rnn | 868 | steps |
| fast trained (cuda graph, gru unroll 16) | ippo_rnn | 531 | steps |
| imitation learning maybe | behavior_cloning | 42 | updates |
| coord_ring seed 373 | ippo_rnn | 373 | steps |
| coord_ring seed 842 | ippo_rnn | 842 | steps |