{
"arch": "x -> Linear(160,H) -> relu -> Linear(H,H) -> relu -> Linear(H,38)",
"batch": 4096,
"data_contested_rows": 77650,
"data_git_sha": "6a4ae961a56f54488441305ce743a98cce9383cc",
"data_rows": 104476,
"data_seed": 1,
"dtype": "f32-le",
"epochs": 300,
"feature_version": 1,
"features_len": 160,
"git_sha": "af792a54efd54faaaff2688538ddfe2a611d03db",
"hidden": 256,
"lr": 0.001,
"param_floats": 116774,
"param_order": [
"l1.weight",
"l1.bias",
"l2.weight",
"l2.bias",
"l3.weight",
"l3.bias"
],
"param_shapes": {
"l1.bias": [
256
],
"l1.weight": [
256,
160
],
"l2.bias": [
256
],
"l2.weight": [
256,
256
],
"l3.bias": [
38
],
"l3.weight": [
38,
256
]
},
"softmax_len": 38,
"teacher": "",
"train_rows": 94028,
"trainer": "pons-trainer 0.1.0",
"val_ce": 0.9673617482185364,
"val_rows": 10448,
"val_top1_constructive": 0.8813944458961487,
"val_top1_contested": 0.7029069066047668,
"val_top1_overall": 0.747990071773529,
"wd": 0.0
}