decision-0.8b / provenance /selection.json
johnsonchromia's picture
Publish Decision-0.8B and verified benchmarks
cb17a40 verified
Raw History Blame Contribute Delete
1.76 kB
{
"protocol_sha256": "96feb75f49f4d90ecc42678153836d5ab1e21de5da68bce8cab96e2123ef71fa",
"selected": {
"adapter": "runs/decision-0.8b-v2/adapter",
"report": "reports/decision-0.8b-v2/decision-development.json",
"step": 9289,
"report_sha256": "701bf9841acbb867d113dff015a52323832deeaeff5b5174be63f1d05f9cf321",
"accuracy": 0.7589686098654709,
"family_macro": 0.7953125
},
"candidates": [
{
"adapter": "runs/decision-0.8b-v1/checkpoint-500",
"report": "reports/decision-0.8b-v1/pilot500-development.json",
"step": 500,
"report_sha256": "ee8903d0b3b4109f0ea84effb5ab7519291e757e47ba97bb8c64a7dc5ca422d7",
"accuracy": 0.6233183856502242,
"family_macro": 0.6508854166666667
},
{
"adapter": "runs/decision-0.8b-v2/checkpoint-1000",
"report": "reports/decision-0.8b-v2/checkpoint1000-development.json",
"step": 1000,
"report_sha256": "4c1e30a22a651d0ac1279d7838ec24a3baf05c4e90635db8322e25cd2bb87917",
"accuracy": 0.6401345291479821,
"family_macro": 0.67140625
},
{
"adapter": "runs/decision-0.8b-v2/checkpoint-5000",
"report": "reports/decision-0.8b-v2/checkpoint5000-development.json",
"step": 5000,
"report_sha256": "8a2b8779d803e90f43f0244525cca94d0e62719a79e1d83c0e3a3dce6e178408",
"accuracy": 0.7432735426008968,
"family_macro": 0.776875
},
{
"adapter": "runs/decision-0.8b-v2/adapter",
"report": "reports/decision-0.8b-v2/decision-development.json",
"step": 9289,
"report_sha256": "701bf9841acbb867d113dff015a52323832deeaeff5b5174be63f1d05f9cf321",
"accuracy": 0.7589686098654709,
"family_macro": 0.7953125
}
],
"test_predictions_used": false
}