ministrike-selfplay-cpu / initial_evaluation.json
BurnyCoder's picture
Publish 32768 transition endpoint and measured evidence
4665e5c verified
Raw History Blame Contribute Delete
8.85 kB
{
"status": "complete",
"source_revision": "74b57fe47b5624d1a470c5ab49516553112f42c6",
"protocol": {
"version": 1,
"policy_identity": "sha256:eb31b8a2f4f8ab7816241219ac4aee369ac4a71ed7437d2cabab1ba8d77eaa29",
"seed_base": 2000000,
"pairs": 4,
"num_envs": 4,
"baselines": [
"random",
"rush",
"guard"
]
},
"seed_base": 2000000,
"pairs": 4,
"num_envs": 4,
"elapsed_seconds": 51.82899999999972,
"rounds": [
{
"pair": "2000000:0",
"baseline": "random",
"side": 0,
"winner": 1,
"score": 0.0,
"seconds": 60.0,
"plants": 0,
"defuses": 0,
"learner_plants": 0,
"learner_defuses": 0,
"damage": 0.0,
"shots": 265,
"inactive_fraction": 0.0
},
{
"pair": "2000000:1",
"baseline": "random",
"side": 0,
"winner": 1,
"score": 0.0,
"seconds": 60.0,
"plants": 0,
"defuses": 0,
"learner_plants": 0,
"learner_defuses": 0,
"damage": 0.0,
"shots": 265,
"inactive_fraction": 0.0
},
{
"pair": "2000000:2",
"baseline": "random",
"side": 0,
"winner": 1,
"score": 0.0,
"seconds": 60.0,
"plants": 0,
"defuses": 0,
"learner_plants": 0,
"learner_defuses": 0,
"damage": 0.0,
"shots": 265,
"inactive_fraction": 0.0
},
{
"pair": "2000000:3",
"baseline": "random",
"side": 0,
"winner": 1,
"score": 0.0,
"seconds": 60.0,
"plants": 0,
"defuses": 0,
"learner_plants": 0,
"learner_defuses": 0,
"damage": 0.0,
"shots": 265,
"inactive_fraction": 0.0
},
{
"pair": "2000000:0",
"baseline": "random",
"side": 1,
"winner": 1,
"score": 1.0,
"seconds": 60.0,
"plants": 0,
"defuses": 0,
"learner_plants": 0,
"learner_defuses": 0,
"damage": 0.0,
"shots": 159,
"inactive_fraction": 0.0
},
{
"pair": "2000000:1",
"baseline": "random",
"side": 1,
"winner": 1,
"score": 1.0,
"seconds": 60.0,
"plants": 0,
"defuses": 0,
"learner_plants": 0,
"learner_defuses": 0,
"damage": 0.0,
"shots": 159,
"inactive_fraction": 0.0
},
{
"pair": "2000000:2",
"baseline": "random",
"side": 1,
"winner": 1,
"score": 1.0,
"seconds": 60.0,
"plants": 0,
"defuses": 0,
"learner_plants": 0,
"learner_defuses": 0,
"damage": 0.0,
"shots": 159,
"inactive_fraction": 0.0
},
{
"pair": "2000000:3",
"baseline": "random",
"side": 1,
"winner": 1,
"score": 1.0,
"seconds": 60.0,
"plants": 0,
"defuses": 0,
"learner_plants": 0,
"learner_defuses": 0,
"damage": 50.0,
"shots": 159,
"inactive_fraction": 0.0
},
{
"pair": "2000000:0",
"baseline": "rush",
"side": 0,
"winner": 1,
"score": 0.0,
"seconds": 60.0,
"plants": 0,
"defuses": 0,
"learner_plants": 0,
"learner_defuses": 0,
"damage": 0.0,
"shots": 265,
"inactive_fraction": 0.0
},
{
"pair": "2000000:1",
"baseline": "rush",
"side": 0,
"winner": 1,
"score": 0.0,
"seconds": 60.0,
"plants": 0,
"defuses": 0,
"learner_plants": 0,
"learner_defuses": 0,
"damage": 0.0,
"shots": 265,
"inactive_fraction": 0.0
},
{
"pair": "2000000:2",
"baseline": "rush",
"side": 0,
"winner": 1,
"score": 0.0,
"seconds": 60.0,
"plants": 0,
"defuses": 0,
"learner_plants": 0,
"learner_defuses": 0,
"damage": 0.0,
"shots": 265,
"inactive_fraction": 0.0
},
{
"pair": "2000000:3",
"baseline": "rush",
"side": 0,
"winner": 1,
"score": 0.0,
"seconds": 60.0,
"plants": 0,
"defuses": 0,
"learner_plants": 0,
"learner_defuses": 0,
"damage": 0.0,
"shots": 265,
"inactive_fraction": 0.0
},
{
"pair": "2000000:2",
"baseline": "rush",
"side": 1,
"winner": 0,
"score": 0.0,
"seconds": 8.25,
"plants": 1,
"defuses": 0,
"learner_plants": 0,
"learner_defuses": 0,
"damage": 0.0,
"shots": 19,
"inactive_fraction": 0.0
},
{
"pair": "2000000:1",
"baseline": "rush",
"side": 1,
"winner": 0,
"score": 0.0,
"seconds": 8.3,
"plants": 1,
"defuses": 0,
"learner_plants": 0,
"learner_defuses": 0,
"damage": 0.0,
"shots": 19,
"inactive_fraction": 0.0
},
{
"pair": "2000000:0",
"baseline": "rush",
"side": 1,
"winner": 0,
"score": 0.0,
"seconds": 32.45,
"plants": 1,
"defuses": 0,
"learner_plants": 0,
"learner_defuses": 0,
"damage": 0.0,
"shots": 40,
"inactive_fraction": 0.0
},
{
"pair": "2000000:3",
"baseline": "rush",
"side": 1,
"winner": 0,
"score": 0.0,
"seconds": 32.45,
"plants": 1,
"defuses": 0,
"learner_plants": 0,
"learner_defuses": 0,
"damage": 0.0,
"shots": 40,
"inactive_fraction": 0.0
},
{
"pair": "2000000:0",
"baseline": "guard",
"side": 0,
"winner": 1,
"score": 0.0,
"seconds": 60.0,
"plants": 0,
"defuses": 0,
"learner_plants": 0,
"learner_defuses": 0,
"damage": 0.0,
"shots": 265,
"inactive_fraction": 0.0
},
{
"pair": "2000000:1",
"baseline": "guard",
"side": 0,
"winner": 1,
"score": 0.0,
"seconds": 60.0,
"plants": 0,
"defuses": 0,
"learner_plants": 0,
"learner_defuses": 0,
"damage": 0.0,
"shots": 265,
"inactive_fraction": 0.0
},
{
"pair": "2000000:2",
"baseline": "guard",
"side": 0,
"winner": 1,
"score": 0.0,
"seconds": 60.0,
"plants": 0,
"defuses": 0,
"learner_plants": 0,
"learner_defuses": 0,
"damage": 0.0,
"shots": 265,
"inactive_fraction": 0.0
},
{
"pair": "2000000:3",
"baseline": "guard",
"side": 0,
"winner": 1,
"score": 0.0,
"seconds": 60.0,
"plants": 0,
"defuses": 0,
"learner_plants": 0,
"learner_defuses": 0,
"damage": 0.0,
"shots": 265,
"inactive_fraction": 0.0
},
{
"pair": "2000000:2",
"baseline": "guard",
"side": 1,
"winner": 0,
"score": 0.0,
"seconds": 33.9,
"plants": 1,
"defuses": 0,
"learner_plants": 0,
"learner_defuses": 0,
"damage": 0.0,
"shots": 38,
"inactive_fraction": 0.0
},
{
"pair": "2000000:1",
"baseline": "guard",
"side": 1,
"winner": 0,
"score": 0.0,
"seconds": 33.95,
"plants": 1,
"defuses": 0,
"learner_plants": 0,
"learner_defuses": 0,
"damage": 0.0,
"shots": 36,
"inactive_fraction": 0.0
},
{
"pair": "2000000:0",
"baseline": "guard",
"side": 1,
"winner": 0,
"score": 0.0,
"seconds": 34.050000000000004,
"plants": 1,
"defuses": 0,
"learner_plants": 0,
"learner_defuses": 0,
"damage": 0.0,
"shots": 38,
"inactive_fraction": 0.0
},
{
"pair": "2000000:3",
"baseline": "guard",
"side": 1,
"winner": 0,
"score": 0.0,
"seconds": 34.1,
"plants": 1,
"defuses": 0,
"learner_plants": 0,
"learner_defuses": 0,
"damage": 0.0,
"shots": 38,
"inactive_fraction": 0.0
}
],
"summary": {
"rounds": 24,
"seed_pairs": 4,
"mean_score": 0.16666666666666666,
"score_ci95": [
0.16666666666666666,
0.16666666666666666
],
"wins": 4,
"draws": 0,
"losses": 20,
"interval_unit": "seed pair, conditional on this checkpoint; not independent training runs",
"by_side": {
"0": 0.0,
"1": 0.3333333333333333
},
"by_baseline": {
"guard": 0.0,
"random": 0.5,
"rush": 0.0
}
}
}