ppo_v8_rust_8gpu/metadata.json

{
  "run_name": "ppo_v8_rust_8gpu",
  "artifact": "policy_update200.ot",
  "latest_alias": "policy_latest.ot",
  "repo": "djmango/openfront-rl",
  "source_pod": "openfront-v8-8gpu-final",
  "source_path": "/root/openfront-ai-v8/rust/checkpoints/policy_update200.ot",
  "uploaded_utc": "2026-07-09T16:27:53Z",
  "git_sha": "d9bdb70ac4e403523aefb00afe7dcce6459689b7",
  "sha256": "37259e1f60b9c8437624108ab6e1269503c13bca1e044bbae8bda9c08b25acb1",
  "size_bytes": 44697096,
  "command": "./target/release/oftrain --num-envs 32 --num-gpus 8 --rollout-len 64 --updates 100000 --epochs 2 --minibatches 4 --device cuda --log-every 1 --ckpt-every 25 --ckpt-dir checkpoints",
  "checkpoint_update": 200,
  "checkpoint_log_line": "[update   200] steps/s=  905.3 decisions_total=  3293184 eps_done=16469 recent_reward=  12.163 pg=-0.0018 v=0.8120 ent=0.204 entq=-0.484 ecoef=0.0096 stage=0 lr=3.00e-4 elapsed=3728s update_s=18.1 collect_s=18.1 train_s=15.5 gpu_mem%=92 min_mean_util%=57 [gpu0=36% gpu1=0% gpu2=32% gpu3=29% gpu4=31% gpu5=14% gpu6=12% gpu7=28%]",
  "last_log_line_at_copy": "[update   211] steps/s=  906.8 decisions_total=  3473408 eps_done=17425 recent_reward=  12.726 pg=-0.0038 v=0.5749 ent=0.231 entq=-0.512 ecoef=0.0096 stage=0 lr=3.00e-4 elapsed=3928s update_s=18.1 collect_s=18.0 train_s=16.0 gpu_mem%=92 min_mean_util%=57 [gpu0=28% gpu1=12% gpu2=33% gpu3=26% gpu4=30% gpu5=0% gpu6=0% gpu7=28%]",
  "format_note": "Rust tch VarStore .ot checkpoint; not a Python torch.save policy.pt. Load with rust/oftrain policy architecture."
}