sam1120's picture
Training in progress, step 120
2615bf1
{
"best_metric": 0.16572275757789612,
"best_model_checkpoint": "/robodata/smodak/Projects/nspl/scripts/terrainseg/training/models/dropoff-utcustom-train-SF-RGBD-b5_6/checkpoint-120",
"epoch": 60.0,
"global_step": 120,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.5,
"learning_rate": 1.6666666666666667e-06,
"loss": 1.2935,
"step": 1
},
{
"epoch": 1.0,
"learning_rate": 3.3333333333333333e-06,
"loss": 1.3048,
"step": 2
},
{
"epoch": 1.5,
"learning_rate": 5e-06,
"loss": 1.2961,
"step": 3
},
{
"epoch": 2.0,
"learning_rate": 6.666666666666667e-06,
"loss": 1.2755,
"step": 4
},
{
"epoch": 2.5,
"learning_rate": 8.333333333333334e-06,
"loss": 1.2586,
"step": 5
},
{
"epoch": 3.0,
"learning_rate": 1e-05,
"loss": 1.246,
"step": 6
},
{
"epoch": 3.5,
"learning_rate": 1.1666666666666668e-05,
"loss": 1.2275,
"step": 7
},
{
"epoch": 4.0,
"learning_rate": 1.3333333333333333e-05,
"loss": 1.2057,
"step": 8
},
{
"epoch": 4.5,
"learning_rate": 1.5000000000000002e-05,
"loss": 1.1914,
"step": 9
},
{
"epoch": 5.0,
"learning_rate": 1.6666666666666667e-05,
"loss": 1.159,
"step": 10
},
{
"epoch": 5.0,
"eval_accuracy_dropoff": 0.5030895204018441,
"eval_accuracy_undropoff": 0.6320608332329993,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.06443018550460354,
"eval_iou_undropoff": 0.6203390578678918,
"eval_iou_unlabeled": 0.0,
"eval_loss": 1.003960132598877,
"eval_mean_accuracy": 0.5675751768174218,
"eval_mean_iou": 0.22825641445749845,
"eval_overall_accuracy": 0.6266983032226563,
"eval_runtime": 2.8797,
"eval_samples_per_second": 6.945,
"eval_steps_per_second": 0.695,
"step": 10
},
{
"epoch": 5.5,
"learning_rate": 1.8333333333333333e-05,
"loss": 1.138,
"step": 11
},
{
"epoch": 6.0,
"learning_rate": 2e-05,
"loss": 1.1,
"step": 12
},
{
"epoch": 6.5,
"learning_rate": 1.9912280701754387e-05,
"loss": 1.0765,
"step": 13
},
{
"epoch": 7.0,
"learning_rate": 1.9824561403508773e-05,
"loss": 1.0466,
"step": 14
},
{
"epoch": 7.5,
"learning_rate": 1.9736842105263158e-05,
"loss": 1.0032,
"step": 15
},
{
"epoch": 8.0,
"learning_rate": 1.9649122807017544e-05,
"loss": 0.9971,
"step": 16
},
{
"epoch": 8.5,
"learning_rate": 1.9561403508771933e-05,
"loss": 0.9459,
"step": 17
},
{
"epoch": 9.0,
"learning_rate": 1.9473684210526318e-05,
"loss": 0.9112,
"step": 18
},
{
"epoch": 9.5,
"learning_rate": 1.9385964912280704e-05,
"loss": 0.8852,
"step": 19
},
{
"epoch": 10.0,
"learning_rate": 1.929824561403509e-05,
"loss": 0.8345,
"step": 20
},
{
"epoch": 10.0,
"eval_accuracy_dropoff": 0.11338333448014863,
"eval_accuracy_undropoff": 0.9505815555977898,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.05546940396815957,
"eval_iou_undropoff": 0.9154202339604258,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.747998058795929,
"eval_mean_accuracy": 0.5319824450389692,
"eval_mean_iou": 0.32362987930952847,
"eval_overall_accuracy": 0.915771484375,
"eval_runtime": 2.8511,
"eval_samples_per_second": 7.015,
"eval_steps_per_second": 0.701,
"step": 20
},
{
"epoch": 10.5,
"learning_rate": 1.9210526315789474e-05,
"loss": 0.8052,
"step": 21
},
{
"epoch": 11.0,
"learning_rate": 1.912280701754386e-05,
"loss": 0.7648,
"step": 22
},
{
"epoch": 11.5,
"learning_rate": 1.9035087719298245e-05,
"loss": 0.7405,
"step": 23
},
{
"epoch": 12.0,
"learning_rate": 1.894736842105263e-05,
"loss": 0.7066,
"step": 24
},
{
"epoch": 12.5,
"learning_rate": 1.885964912280702e-05,
"loss": 0.6675,
"step": 25
},
{
"epoch": 13.0,
"learning_rate": 1.8771929824561405e-05,
"loss": 0.6358,
"step": 26
},
{
"epoch": 13.5,
"learning_rate": 1.868421052631579e-05,
"loss": 0.6127,
"step": 27
},
{
"epoch": 14.0,
"learning_rate": 1.8596491228070176e-05,
"loss": 0.5935,
"step": 28
},
{
"epoch": 14.5,
"learning_rate": 1.8508771929824562e-05,
"loss": 0.5549,
"step": 29
},
{
"epoch": 15.0,
"learning_rate": 1.8421052631578947e-05,
"loss": 0.5406,
"step": 30
},
{
"epoch": 15.0,
"eval_accuracy_dropoff": 0.017945365719397234,
"eval_accuracy_undropoff": 0.9917655826949273,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.015735932390196418,
"eval_iou_undropoff": 0.9512822485309885,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.5476881265640259,
"eval_mean_accuracy": 0.5048554742071623,
"eval_mean_iou": 0.322339393640395,
"eval_overall_accuracy": 0.9512748718261719,
"eval_runtime": 5.6484,
"eval_samples_per_second": 3.541,
"eval_steps_per_second": 0.354,
"step": 30
},
{
"epoch": 15.5,
"learning_rate": 1.8333333333333333e-05,
"loss": 0.53,
"step": 31
},
{
"epoch": 16.0,
"learning_rate": 1.824561403508772e-05,
"loss": 0.5004,
"step": 32
},
{
"epoch": 16.5,
"learning_rate": 1.8157894736842107e-05,
"loss": 0.4965,
"step": 33
},
{
"epoch": 17.0,
"learning_rate": 1.8070175438596493e-05,
"loss": 0.4653,
"step": 34
},
{
"epoch": 17.5,
"learning_rate": 1.7982456140350878e-05,
"loss": 0.4434,
"step": 35
},
{
"epoch": 18.0,
"learning_rate": 1.7894736842105264e-05,
"loss": 0.4414,
"step": 36
},
{
"epoch": 18.5,
"learning_rate": 1.780701754385965e-05,
"loss": 0.3915,
"step": 37
},
{
"epoch": 19.0,
"learning_rate": 1.7719298245614035e-05,
"loss": 0.4001,
"step": 38
},
{
"epoch": 19.5,
"learning_rate": 1.763157894736842e-05,
"loss": 0.4204,
"step": 39
},
{
"epoch": 20.0,
"learning_rate": 1.754385964912281e-05,
"loss": 0.3695,
"step": 40
},
{
"epoch": 20.0,
"eval_accuracy_dropoff": 0.014605839583476686,
"eval_accuracy_undropoff": 0.9925813227566402,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.012491320023381993,
"eval_iou_undropoff": 0.951889892701647,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.45899271965026855,
"eval_mean_accuracy": 0.5035935811700585,
"eval_mean_iou": 0.3214604042416764,
"eval_overall_accuracy": 0.9519178390502929,
"eval_runtime": 3.3683,
"eval_samples_per_second": 5.938,
"eval_steps_per_second": 0.594,
"step": 40
},
{
"epoch": 20.5,
"learning_rate": 1.7456140350877195e-05,
"loss": 0.3526,
"step": 41
},
{
"epoch": 21.0,
"learning_rate": 1.736842105263158e-05,
"loss": 0.3346,
"step": 42
},
{
"epoch": 21.5,
"learning_rate": 1.7280701754385966e-05,
"loss": 0.3645,
"step": 43
},
{
"epoch": 22.0,
"learning_rate": 1.719298245614035e-05,
"loss": 0.3017,
"step": 44
},
{
"epoch": 22.5,
"learning_rate": 1.7105263157894737e-05,
"loss": 0.3384,
"step": 45
},
{
"epoch": 23.0,
"learning_rate": 1.7017543859649125e-05,
"loss": 0.3197,
"step": 46
},
{
"epoch": 23.5,
"learning_rate": 1.692982456140351e-05,
"loss": 0.3031,
"step": 47
},
{
"epoch": 24.0,
"learning_rate": 1.6842105263157896e-05,
"loss": 0.3068,
"step": 48
},
{
"epoch": 24.5,
"learning_rate": 1.6754385964912282e-05,
"loss": 0.2974,
"step": 49
},
{
"epoch": 25.0,
"learning_rate": 1.6666666666666667e-05,
"loss": 0.3053,
"step": 50
},
{
"epoch": 25.0,
"eval_accuracy_dropoff": 0.002344090460790385,
"eval_accuracy_undropoff": 0.9979102009299715,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.002236568552358026,
"eval_iou_undropoff": 0.9565110735242025,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.3790256679058075,
"eval_mean_accuracy": 0.500127145695381,
"eval_mean_iou": 0.3195825473588535,
"eval_overall_accuracy": 0.9565153121948242,
"eval_runtime": 2.7928,
"eval_samples_per_second": 7.161,
"eval_steps_per_second": 0.716,
"step": 50
},
{
"epoch": 25.5,
"learning_rate": 1.6578947368421053e-05,
"loss": 0.2815,
"step": 51
},
{
"epoch": 26.0,
"learning_rate": 1.649122807017544e-05,
"loss": 0.2835,
"step": 52
},
{
"epoch": 26.5,
"learning_rate": 1.6403508771929827e-05,
"loss": 0.2771,
"step": 53
},
{
"epoch": 27.0,
"learning_rate": 1.6315789473684213e-05,
"loss": 0.2843,
"step": 54
},
{
"epoch": 27.5,
"learning_rate": 1.62280701754386e-05,
"loss": 0.2831,
"step": 55
},
{
"epoch": 28.0,
"learning_rate": 1.6140350877192984e-05,
"loss": 0.2437,
"step": 56
},
{
"epoch": 28.5,
"learning_rate": 1.605263157894737e-05,
"loss": 0.2769,
"step": 57
},
{
"epoch": 29.0,
"learning_rate": 1.5964912280701755e-05,
"loss": 0.2448,
"step": 58
},
{
"epoch": 29.5,
"learning_rate": 1.5877192982456144e-05,
"loss": 0.2692,
"step": 59
},
{
"epoch": 30.0,
"learning_rate": 1.578947368421053e-05,
"loss": 0.2436,
"step": 60
},
{
"epoch": 30.0,
"eval_accuracy_dropoff": 0.00587169430491525,
"eval_accuracy_undropoff": 0.9980554778865586,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.0056198027800461875,
"eval_iou_undropoff": 0.9567906746031746,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.3303200304508209,
"eval_mean_accuracy": 0.5019635860957369,
"eval_mean_iou": 0.48120523869161036,
"eval_overall_accuracy": 0.9568012237548829,
"eval_runtime": 2.7601,
"eval_samples_per_second": 7.246,
"eval_steps_per_second": 0.725,
"step": 60
},
{
"epoch": 30.5,
"learning_rate": 1.5701754385964915e-05,
"loss": 0.2546,
"step": 61
},
{
"epoch": 31.0,
"learning_rate": 1.56140350877193e-05,
"loss": 0.2305,
"step": 62
},
{
"epoch": 31.5,
"learning_rate": 1.5526315789473686e-05,
"loss": 0.2386,
"step": 63
},
{
"epoch": 32.0,
"learning_rate": 1.543859649122807e-05,
"loss": 0.2412,
"step": 64
},
{
"epoch": 32.5,
"learning_rate": 1.535087719298246e-05,
"loss": 0.2459,
"step": 65
},
{
"epoch": 33.0,
"learning_rate": 1.5263157894736846e-05,
"loss": 0.2444,
"step": 66
},
{
"epoch": 33.5,
"learning_rate": 1.517543859649123e-05,
"loss": 0.2415,
"step": 67
},
{
"epoch": 34.0,
"learning_rate": 1.5087719298245615e-05,
"loss": 0.2233,
"step": 68
},
{
"epoch": 34.5,
"learning_rate": 1.5000000000000002e-05,
"loss": 0.2269,
"step": 69
},
{
"epoch": 35.0,
"learning_rate": 1.4912280701754388e-05,
"loss": 0.2148,
"step": 70
},
{
"epoch": 35.0,
"eval_accuracy_dropoff": 0.0008394687951558522,
"eval_accuracy_undropoff": 0.9995707364447146,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.0008312438678731058,
"eval_iou_undropoff": 0.9580427783638841,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.2739110589027405,
"eval_mean_accuracy": 0.5002051026199352,
"eval_mean_iou": 0.4794370111158786,
"eval_overall_accuracy": 0.9580442428588867,
"eval_runtime": 2.7182,
"eval_samples_per_second": 7.358,
"eval_steps_per_second": 0.736,
"step": 70
},
{
"epoch": 35.5,
"learning_rate": 1.4824561403508773e-05,
"loss": 0.2208,
"step": 71
},
{
"epoch": 36.0,
"learning_rate": 1.4736842105263159e-05,
"loss": 0.2169,
"step": 72
},
{
"epoch": 36.5,
"learning_rate": 1.4649122807017544e-05,
"loss": 0.2418,
"step": 73
},
{
"epoch": 37.0,
"learning_rate": 1.4561403508771931e-05,
"loss": 0.1911,
"step": 74
},
{
"epoch": 37.5,
"learning_rate": 1.4473684210526317e-05,
"loss": 0.2089,
"step": 75
},
{
"epoch": 38.0,
"learning_rate": 1.4385964912280704e-05,
"loss": 0.2248,
"step": 76
},
{
"epoch": 38.5,
"learning_rate": 1.429824561403509e-05,
"loss": 0.2164,
"step": 77
},
{
"epoch": 39.0,
"learning_rate": 1.4210526315789475e-05,
"loss": 0.2224,
"step": 78
},
{
"epoch": 39.5,
"learning_rate": 1.412280701754386e-05,
"loss": 0.2239,
"step": 79
},
{
"epoch": 40.0,
"learning_rate": 1.4035087719298246e-05,
"loss": 0.1983,
"step": 80
},
{
"epoch": 40.0,
"eval_accuracy_dropoff": 0.05823528062570243,
"eval_accuracy_undropoff": 0.998591012530635,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.0564034210818616,
"eval_iou_undropoff": 0.9593934057781895,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.23478250205516815,
"eval_mean_accuracy": 0.5284131465781687,
"eval_mean_iou": 0.5078984134300256,
"eval_overall_accuracy": 0.9594917297363281,
"eval_runtime": 2.9691,
"eval_samples_per_second": 6.736,
"eval_steps_per_second": 0.674,
"step": 80
},
{
"epoch": 40.5,
"learning_rate": 1.3947368421052631e-05,
"loss": 0.2082,
"step": 81
},
{
"epoch": 41.0,
"learning_rate": 1.385964912280702e-05,
"loss": 0.1867,
"step": 82
},
{
"epoch": 41.5,
"learning_rate": 1.3771929824561406e-05,
"loss": 0.189,
"step": 83
},
{
"epoch": 42.0,
"learning_rate": 1.3684210526315791e-05,
"loss": 0.2007,
"step": 84
},
{
"epoch": 42.5,
"learning_rate": 1.3596491228070177e-05,
"loss": 0.1998,
"step": 85
},
{
"epoch": 43.0,
"learning_rate": 1.3508771929824562e-05,
"loss": 0.1833,
"step": 86
},
{
"epoch": 43.5,
"learning_rate": 1.3421052631578948e-05,
"loss": 0.2017,
"step": 87
},
{
"epoch": 44.0,
"learning_rate": 1.3333333333333333e-05,
"loss": 0.1757,
"step": 88
},
{
"epoch": 44.5,
"learning_rate": 1.3245614035087719e-05,
"loss": 0.1922,
"step": 89
},
{
"epoch": 45.0,
"learning_rate": 1.3157894736842108e-05,
"loss": 0.1784,
"step": 90
},
{
"epoch": 45.0,
"eval_accuracy_dropoff": 0.29602513819124293,
"eval_accuracy_undropoff": 0.9920426039600906,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.2501434219706954,
"eval_iou_undropoff": 0.9626429123728263,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.21779485046863556,
"eval_mean_accuracy": 0.6440338710756668,
"eval_mean_iou": 0.6063931671717608,
"eval_overall_accuracy": 0.9631027221679688,
"eval_runtime": 2.7873,
"eval_samples_per_second": 7.175,
"eval_steps_per_second": 0.718,
"step": 90
},
{
"epoch": 45.5,
"learning_rate": 1.3070175438596493e-05,
"loss": 0.184,
"step": 91
},
{
"epoch": 46.0,
"learning_rate": 1.2982456140350879e-05,
"loss": 0.1835,
"step": 92
},
{
"epoch": 46.5,
"learning_rate": 1.2894736842105264e-05,
"loss": 0.1654,
"step": 93
},
{
"epoch": 47.0,
"learning_rate": 1.280701754385965e-05,
"loss": 0.181,
"step": 94
},
{
"epoch": 47.5,
"learning_rate": 1.2719298245614035e-05,
"loss": 0.1782,
"step": 95
},
{
"epoch": 48.0,
"learning_rate": 1.263157894736842e-05,
"loss": 0.172,
"step": 96
},
{
"epoch": 48.5,
"learning_rate": 1.2543859649122808e-05,
"loss": 0.1819,
"step": 97
},
{
"epoch": 49.0,
"learning_rate": 1.2456140350877195e-05,
"loss": 0.1685,
"step": 98
},
{
"epoch": 49.5,
"learning_rate": 1.236842105263158e-05,
"loss": 0.1616,
"step": 99
},
{
"epoch": 50.0,
"learning_rate": 1.2280701754385966e-05,
"loss": 0.1631,
"step": 100
},
{
"epoch": 50.0,
"eval_accuracy_dropoff": 0.3760499094015918,
"eval_accuracy_undropoff": 0.9860599794821175,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.2846008568194916,
"eval_iou_undropoff": 0.9600719098963882,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.1942574679851532,
"eval_mean_accuracy": 0.6810549444418547,
"eval_mean_iou": 0.62233638335794,
"eval_overall_accuracy": 0.9606962203979492,
"eval_runtime": 2.9092,
"eval_samples_per_second": 6.875,
"eval_steps_per_second": 0.687,
"step": 100
},
{
"epoch": 50.5,
"learning_rate": 1.2192982456140352e-05,
"loss": 0.153,
"step": 101
},
{
"epoch": 51.0,
"learning_rate": 1.2105263157894737e-05,
"loss": 0.172,
"step": 102
},
{
"epoch": 51.5,
"learning_rate": 1.2017543859649124e-05,
"loss": 0.1695,
"step": 103
},
{
"epoch": 52.0,
"learning_rate": 1.192982456140351e-05,
"loss": 0.1743,
"step": 104
},
{
"epoch": 52.5,
"learning_rate": 1.1842105263157895e-05,
"loss": 0.1628,
"step": 105
},
{
"epoch": 53.0,
"learning_rate": 1.1754385964912282e-05,
"loss": 0.1585,
"step": 106
},
{
"epoch": 53.5,
"learning_rate": 1.1666666666666668e-05,
"loss": 0.1494,
"step": 107
},
{
"epoch": 54.0,
"learning_rate": 1.1578947368421053e-05,
"loss": 0.15,
"step": 108
},
{
"epoch": 54.5,
"learning_rate": 1.1491228070175439e-05,
"loss": 0.1556,
"step": 109
},
{
"epoch": 55.0,
"learning_rate": 1.1403508771929826e-05,
"loss": 0.1468,
"step": 110
},
{
"epoch": 55.0,
"eval_accuracy_dropoff": 0.358338494002156,
"eval_accuracy_undropoff": 0.9878841008301682,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.2801101564853196,
"eval_iou_undropoff": 0.9611289112145298,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.17585495114326477,
"eval_mean_accuracy": 0.6731112974161622,
"eval_mean_iou": 0.6206195338499247,
"eval_overall_accuracy": 0.9617080688476562,
"eval_runtime": 5.2724,
"eval_samples_per_second": 3.793,
"eval_steps_per_second": 0.379,
"step": 110
},
{
"epoch": 55.5,
"learning_rate": 1.1315789473684212e-05,
"loss": 0.1559,
"step": 111
},
{
"epoch": 56.0,
"learning_rate": 1.1228070175438597e-05,
"loss": 0.1603,
"step": 112
},
{
"epoch": 56.5,
"learning_rate": 1.1140350877192983e-05,
"loss": 0.1488,
"step": 113
},
{
"epoch": 57.0,
"learning_rate": 1.105263157894737e-05,
"loss": 0.1414,
"step": 114
},
{
"epoch": 57.5,
"learning_rate": 1.0964912280701755e-05,
"loss": 0.1461,
"step": 115
},
{
"epoch": 58.0,
"learning_rate": 1.0877192982456142e-05,
"loss": 0.1519,
"step": 116
},
{
"epoch": 58.5,
"learning_rate": 1.0789473684210528e-05,
"loss": 0.1509,
"step": 117
},
{
"epoch": 59.0,
"learning_rate": 1.0701754385964913e-05,
"loss": 0.1393,
"step": 118
},
{
"epoch": 59.5,
"learning_rate": 1.0614035087719299e-05,
"loss": 0.1369,
"step": 119
},
{
"epoch": 60.0,
"learning_rate": 1.0526315789473684e-05,
"loss": 0.1353,
"step": 120
},
{
"epoch": 60.0,
"eval_accuracy_dropoff": 0.2731347049244249,
"eval_accuracy_undropoff": 0.9938581678983698,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.23926190543163342,
"eval_iou_undropoff": 0.9634762386709105,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.16572275757789612,
"eval_mean_accuracy": 0.6334964364113973,
"eval_mean_iou": 0.601369072051272,
"eval_overall_accuracy": 0.9638910293579102,
"eval_runtime": 2.7535,
"eval_samples_per_second": 7.264,
"eval_steps_per_second": 0.726,
"step": 120
}
],
"max_steps": 240,
"num_train_epochs": 120,
"total_flos": 1.3892743058581094e+18,
"trial_name": null,
"trial_params": null
}