sam1120's picture
Training in progress, step 240
2c85a5b
{
"best_metric": 0.1285596638917923,
"best_model_checkpoint": "/robodata/smodak/Projects/nspl/scripts/terrainseg/training/models/dropoff-utcustom-train-SF-RGBD-b5_7/checkpoint-160",
"epoch": 120.0,
"global_step": 240,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.5,
"learning_rate": 4.166666666666667e-06,
"loss": 1.1395,
"step": 1
},
{
"epoch": 1.0,
"learning_rate": 8.333333333333334e-06,
"loss": 1.1418,
"step": 2
},
{
"epoch": 1.5,
"learning_rate": 1.25e-05,
"loss": 1.1243,
"step": 3
},
{
"epoch": 2.0,
"learning_rate": 1.6666666666666667e-05,
"loss": 1.1018,
"step": 4
},
{
"epoch": 2.5,
"learning_rate": 2.0833333333333336e-05,
"loss": 1.0687,
"step": 5
},
{
"epoch": 3.0,
"learning_rate": 2.5e-05,
"loss": 1.0541,
"step": 6
},
{
"epoch": 3.5,
"learning_rate": 2.916666666666667e-05,
"loss": 1.022,
"step": 7
},
{
"epoch": 4.0,
"learning_rate": 3.3333333333333335e-05,
"loss": 0.9945,
"step": 8
},
{
"epoch": 4.5,
"learning_rate": 3.7500000000000003e-05,
"loss": 0.9642,
"step": 9
},
{
"epoch": 5.0,
"learning_rate": 4.166666666666667e-05,
"loss": 0.9278,
"step": 10
},
{
"epoch": 5.0,
"eval_accuracy_dropoff": 0.2009128649739673,
"eval_accuracy_undropoff": 0.9081783961224983,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.0807428683093736,
"eval_iou_undropoff": 0.8784609129926589,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.845426082611084,
"eval_mean_accuracy": 0.5545456305482328,
"eval_mean_iou": 0.3197345937673442,
"eval_overall_accuracy": 0.8787708282470703,
"eval_runtime": 2.7556,
"eval_samples_per_second": 7.258,
"eval_steps_per_second": 0.726,
"step": 10
},
{
"epoch": 5.5,
"learning_rate": 4.5833333333333334e-05,
"loss": 0.8786,
"step": 11
},
{
"epoch": 6.0,
"learning_rate": 5e-05,
"loss": 0.8476,
"step": 12
},
{
"epoch": 6.5,
"learning_rate": 4.9780701754385966e-05,
"loss": 0.7993,
"step": 13
},
{
"epoch": 7.0,
"learning_rate": 4.956140350877193e-05,
"loss": 0.7682,
"step": 14
},
{
"epoch": 7.5,
"learning_rate": 4.9342105263157894e-05,
"loss": 0.7194,
"step": 15
},
{
"epoch": 8.0,
"learning_rate": 4.912280701754386e-05,
"loss": 0.7016,
"step": 16
},
{
"epoch": 8.5,
"learning_rate": 4.890350877192983e-05,
"loss": 0.6592,
"step": 17
},
{
"epoch": 9.0,
"learning_rate": 4.868421052631579e-05,
"loss": 0.6162,
"step": 18
},
{
"epoch": 9.5,
"learning_rate": 4.8464912280701755e-05,
"loss": 0.5957,
"step": 19
},
{
"epoch": 10.0,
"learning_rate": 4.824561403508772e-05,
"loss": 0.5551,
"step": 20
},
{
"epoch": 10.0,
"eval_accuracy_dropoff": 0.013546182251886512,
"eval_accuracy_undropoff": 0.994846449222221,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.0122168669714333,
"eval_iou_undropoff": 0.9540401274092,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.4667937755584717,
"eval_mean_accuracy": 0.5041963157370537,
"eval_mean_iou": 0.32208566479354445,
"eval_overall_accuracy": 0.9540447235107422,
"eval_runtime": 2.9472,
"eval_samples_per_second": 6.786,
"eval_steps_per_second": 0.679,
"step": 20
},
{
"epoch": 10.5,
"learning_rate": 4.802631578947368e-05,
"loss": 0.5393,
"step": 21
},
{
"epoch": 11.0,
"learning_rate": 4.780701754385965e-05,
"loss": 0.5054,
"step": 22
},
{
"epoch": 11.5,
"learning_rate": 4.758771929824562e-05,
"loss": 0.4768,
"step": 23
},
{
"epoch": 12.0,
"learning_rate": 4.736842105263158e-05,
"loss": 0.4747,
"step": 24
},
{
"epoch": 12.5,
"learning_rate": 4.714912280701755e-05,
"loss": 0.4453,
"step": 25
},
{
"epoch": 13.0,
"learning_rate": 4.6929824561403515e-05,
"loss": 0.4263,
"step": 26
},
{
"epoch": 13.5,
"learning_rate": 4.671052631578948e-05,
"loss": 0.4107,
"step": 27
},
{
"epoch": 14.0,
"learning_rate": 4.649122807017544e-05,
"loss": 0.3888,
"step": 28
},
{
"epoch": 14.5,
"learning_rate": 4.6271929824561406e-05,
"loss": 0.3656,
"step": 29
},
{
"epoch": 15.0,
"learning_rate": 4.605263157894737e-05,
"loss": 0.3667,
"step": 30
},
{
"epoch": 15.0,
"eval_accuracy_dropoff": 0.008839652285602882,
"eval_accuracy_undropoff": 0.9981533905751077,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.0085100181505836,
"eval_iou_undropoff": 0.9570052236822447,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.33538389205932617,
"eval_mean_accuracy": 0.5034965214303553,
"eval_mean_iou": 0.3218384139442761,
"eval_overall_accuracy": 0.9570184707641601,
"eval_runtime": 11.3505,
"eval_samples_per_second": 1.762,
"eval_steps_per_second": 0.176,
"step": 30
},
{
"epoch": 15.5,
"learning_rate": 4.5833333333333334e-05,
"loss": 0.3391,
"step": 31
},
{
"epoch": 16.0,
"learning_rate": 4.56140350877193e-05,
"loss": 0.3252,
"step": 32
},
{
"epoch": 16.5,
"learning_rate": 4.539473684210527e-05,
"loss": 0.3199,
"step": 33
},
{
"epoch": 17.0,
"learning_rate": 4.517543859649123e-05,
"loss": 0.2957,
"step": 34
},
{
"epoch": 17.5,
"learning_rate": 4.4956140350877196e-05,
"loss": 0.2837,
"step": 35
},
{
"epoch": 18.0,
"learning_rate": 4.473684210526316e-05,
"loss": 0.2764,
"step": 36
},
{
"epoch": 18.5,
"learning_rate": 4.451754385964912e-05,
"loss": 0.239,
"step": 37
},
{
"epoch": 19.0,
"learning_rate": 4.429824561403509e-05,
"loss": 0.2586,
"step": 38
},
{
"epoch": 19.5,
"learning_rate": 4.407894736842105e-05,
"loss": 0.2539,
"step": 39
},
{
"epoch": 20.0,
"learning_rate": 4.3859649122807014e-05,
"loss": 0.2402,
"step": 40
},
{
"epoch": 20.0,
"eval_accuracy_dropoff": 0.31156219179338973,
"eval_accuracy_undropoff": 0.9867716375598645,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.23875964087096524,
"eval_iou_undropoff": 0.9581548537274001,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.26776206493377686,
"eval_mean_accuracy": 0.6491669146766271,
"eval_mean_iou": 0.5984572472991827,
"eval_overall_accuracy": 0.9586969375610351,
"eval_runtime": 3.4868,
"eval_samples_per_second": 5.736,
"eval_steps_per_second": 0.574,
"step": 40
},
{
"epoch": 20.5,
"learning_rate": 4.3640350877192985e-05,
"loss": 0.2083,
"step": 41
},
{
"epoch": 21.0,
"learning_rate": 4.342105263157895e-05,
"loss": 0.1978,
"step": 42
},
{
"epoch": 21.5,
"learning_rate": 4.320175438596491e-05,
"loss": 0.2042,
"step": 43
},
{
"epoch": 22.0,
"learning_rate": 4.298245614035088e-05,
"loss": 0.1683,
"step": 44
},
{
"epoch": 22.5,
"learning_rate": 4.2763157894736847e-05,
"loss": 0.1905,
"step": 45
},
{
"epoch": 23.0,
"learning_rate": 4.254385964912281e-05,
"loss": 0.1825,
"step": 46
},
{
"epoch": 23.5,
"learning_rate": 4.2324561403508774e-05,
"loss": 0.1595,
"step": 47
},
{
"epoch": 24.0,
"learning_rate": 4.210526315789474e-05,
"loss": 0.1808,
"step": 48
},
{
"epoch": 24.5,
"learning_rate": 4.188596491228071e-05,
"loss": 0.1612,
"step": 49
},
{
"epoch": 25.0,
"learning_rate": 4.166666666666667e-05,
"loss": 0.1562,
"step": 50
},
{
"epoch": 25.0,
"eval_accuracy_dropoff": 0.35438886212986537,
"eval_accuracy_undropoff": 0.9895070235438224,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.2853675924660444,
"eval_iou_undropoff": 0.9625474168179068,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.2101479470729828,
"eval_mean_accuracy": 0.6719479428368439,
"eval_mean_iou": 0.6239575046419756,
"eval_overall_accuracy": 0.9630992889404297,
"eval_runtime": 2.857,
"eval_samples_per_second": 7.0,
"eval_steps_per_second": 0.7,
"step": 50
},
{
"epoch": 25.5,
"learning_rate": 4.1447368421052636e-05,
"loss": 0.1405,
"step": 51
},
{
"epoch": 26.0,
"learning_rate": 4.12280701754386e-05,
"loss": 0.1564,
"step": 52
},
{
"epoch": 26.5,
"learning_rate": 4.100877192982456e-05,
"loss": 0.1386,
"step": 53
},
{
"epoch": 27.0,
"learning_rate": 4.078947368421053e-05,
"loss": 0.1369,
"step": 54
},
{
"epoch": 27.5,
"learning_rate": 4.057017543859649e-05,
"loss": 0.1337,
"step": 55
},
{
"epoch": 28.0,
"learning_rate": 4.0350877192982455e-05,
"loss": 0.1216,
"step": 56
},
{
"epoch": 28.5,
"learning_rate": 4.0131578947368425e-05,
"loss": 0.1219,
"step": 57
},
{
"epoch": 29.0,
"learning_rate": 3.991228070175439e-05,
"loss": 0.1291,
"step": 58
},
{
"epoch": 29.5,
"learning_rate": 3.969298245614035e-05,
"loss": 0.1128,
"step": 59
},
{
"epoch": 30.0,
"learning_rate": 3.9473684210526316e-05,
"loss": 0.1159,
"step": 60
},
{
"epoch": 30.0,
"eval_accuracy_dropoff": 0.3353287919447694,
"eval_accuracy_undropoff": 0.9927785412004454,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.28747610920159505,
"eval_iou_undropoff": 0.9649536343906318,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.17043346166610718,
"eval_mean_accuracy": 0.6640536665726073,
"eval_mean_iou": 0.6262148717961133,
"eval_overall_accuracy": 0.9654422760009765,
"eval_runtime": 2.9147,
"eval_samples_per_second": 6.862,
"eval_steps_per_second": 0.686,
"step": 60
},
{
"epoch": 30.5,
"learning_rate": 3.925438596491228e-05,
"loss": 0.1099,
"step": 61
},
{
"epoch": 31.0,
"learning_rate": 3.9035087719298244e-05,
"loss": 0.1059,
"step": 62
},
{
"epoch": 31.5,
"learning_rate": 3.8815789473684214e-05,
"loss": 0.1125,
"step": 63
},
{
"epoch": 32.0,
"learning_rate": 3.859649122807018e-05,
"loss": 0.1054,
"step": 64
},
{
"epoch": 32.5,
"learning_rate": 3.837719298245614e-05,
"loss": 0.0973,
"step": 65
},
{
"epoch": 33.0,
"learning_rate": 3.815789473684211e-05,
"loss": 0.1075,
"step": 66
},
{
"epoch": 33.5,
"learning_rate": 3.7938596491228076e-05,
"loss": 0.0956,
"step": 67
},
{
"epoch": 34.0,
"learning_rate": 3.771929824561404e-05,
"loss": 0.0914,
"step": 68
},
{
"epoch": 34.5,
"learning_rate": 3.7500000000000003e-05,
"loss": 0.0927,
"step": 69
},
{
"epoch": 35.0,
"learning_rate": 3.728070175438597e-05,
"loss": 0.0869,
"step": 70
},
{
"epoch": 35.0,
"eval_accuracy_dropoff": 0.3719947705222597,
"eval_accuracy_undropoff": 0.9914632872195085,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.3108309600561154,
"eval_iou_undropoff": 0.9651674894760284,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.144324392080307,
"eval_mean_accuracy": 0.6817290288708842,
"eval_mean_iou": 0.6379992247660718,
"eval_overall_accuracy": 0.9657062530517578,
"eval_runtime": 2.8546,
"eval_samples_per_second": 7.006,
"eval_steps_per_second": 0.701,
"step": 70
},
{
"epoch": 35.5,
"learning_rate": 3.706140350877193e-05,
"loss": 0.0825,
"step": 71
},
{
"epoch": 36.0,
"learning_rate": 3.6842105263157895e-05,
"loss": 0.0889,
"step": 72
},
{
"epoch": 36.5,
"learning_rate": 3.662280701754386e-05,
"loss": 0.09,
"step": 73
},
{
"epoch": 37.0,
"learning_rate": 3.640350877192983e-05,
"loss": 0.0929,
"step": 74
},
{
"epoch": 37.5,
"learning_rate": 3.618421052631579e-05,
"loss": 0.0851,
"step": 75
},
{
"epoch": 38.0,
"learning_rate": 3.5964912280701756e-05,
"loss": 0.0811,
"step": 76
},
{
"epoch": 38.5,
"learning_rate": 3.574561403508772e-05,
"loss": 0.0817,
"step": 77
},
{
"epoch": 39.0,
"learning_rate": 3.5526315789473684e-05,
"loss": 0.0867,
"step": 78
},
{
"epoch": 39.5,
"learning_rate": 3.530701754385965e-05,
"loss": 0.0818,
"step": 79
},
{
"epoch": 40.0,
"learning_rate": 3.508771929824561e-05,
"loss": 0.079,
"step": 80
},
{
"epoch": 40.0,
"eval_accuracy_dropoff": 0.2766485469850226,
"eval_accuracy_undropoff": 0.995262379139025,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.24941170632048668,
"eval_iou_undropoff": 0.9649801295572932,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.13497576117515564,
"eval_mean_accuracy": 0.6359554630620239,
"eval_mean_iou": 0.6071959179388899,
"eval_overall_accuracy": 0.9653829574584961,
"eval_runtime": 2.8312,
"eval_samples_per_second": 7.064,
"eval_steps_per_second": 0.706,
"step": 80
},
{
"epoch": 40.5,
"learning_rate": 3.4868421052631575e-05,
"loss": 0.073,
"step": 81
},
{
"epoch": 41.0,
"learning_rate": 3.4649122807017546e-05,
"loss": 0.0825,
"step": 82
},
{
"epoch": 41.5,
"learning_rate": 3.442982456140351e-05,
"loss": 0.0688,
"step": 83
},
{
"epoch": 42.0,
"learning_rate": 3.421052631578947e-05,
"loss": 0.0813,
"step": 84
},
{
"epoch": 42.5,
"learning_rate": 3.3991228070175444e-05,
"loss": 0.0745,
"step": 85
},
{
"epoch": 43.0,
"learning_rate": 3.377192982456141e-05,
"loss": 0.0759,
"step": 86
},
{
"epoch": 43.5,
"learning_rate": 3.355263157894737e-05,
"loss": 0.0706,
"step": 87
},
{
"epoch": 44.0,
"learning_rate": 3.3333333333333335e-05,
"loss": 0.0767,
"step": 88
},
{
"epoch": 44.5,
"learning_rate": 3.31140350877193e-05,
"loss": 0.0736,
"step": 89
},
{
"epoch": 45.0,
"learning_rate": 3.289473684210527e-05,
"loss": 0.0647,
"step": 90
},
{
"epoch": 45.0,
"eval_accuracy_dropoff": 0.20900479368792863,
"eval_accuracy_undropoff": 0.9971020232303824,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.195917559996044,
"eval_iou_undropoff": 0.9640208661467319,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.13696631789207458,
"eval_mean_accuracy": 0.6030534084591556,
"eval_mean_iou": 0.5799692130713879,
"eval_overall_accuracy": 0.9643335342407227,
"eval_runtime": 2.9184,
"eval_samples_per_second": 6.853,
"eval_steps_per_second": 0.685,
"step": 90
},
{
"epoch": 45.5,
"learning_rate": 3.267543859649123e-05,
"loss": 0.0677,
"step": 91
},
{
"epoch": 46.0,
"learning_rate": 3.24561403508772e-05,
"loss": 0.0656,
"step": 92
},
{
"epoch": 46.5,
"learning_rate": 3.223684210526316e-05,
"loss": 0.0811,
"step": 93
},
{
"epoch": 47.0,
"learning_rate": 3.2017543859649124e-05,
"loss": 0.062,
"step": 94
},
{
"epoch": 47.5,
"learning_rate": 3.179824561403509e-05,
"loss": 0.0664,
"step": 95
},
{
"epoch": 48.0,
"learning_rate": 3.157894736842105e-05,
"loss": 0.0622,
"step": 96
},
{
"epoch": 48.5,
"learning_rate": 3.1359649122807015e-05,
"loss": 0.0658,
"step": 97
},
{
"epoch": 49.0,
"learning_rate": 3.1140350877192986e-05,
"loss": 0.0667,
"step": 98
},
{
"epoch": 49.5,
"learning_rate": 3.092105263157895e-05,
"loss": 0.0632,
"step": 99
},
{
"epoch": 50.0,
"learning_rate": 3.0701754385964913e-05,
"loss": 0.0587,
"step": 100
},
{
"epoch": 50.0,
"eval_accuracy_dropoff": 0.37065987752012663,
"eval_accuracy_undropoff": 0.9884823234760597,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.29289882915866167,
"eval_iou_undropoff": 0.9622113420215658,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.1336326152086258,
"eval_mean_accuracy": 0.6795711004980931,
"eval_mean_iou": 0.6275550855901137,
"eval_overall_accuracy": 0.9627937316894531,
"eval_runtime": 2.7754,
"eval_samples_per_second": 7.206,
"eval_steps_per_second": 0.721,
"step": 100
},
{
"epoch": 50.5,
"learning_rate": 3.048245614035088e-05,
"loss": 0.0589,
"step": 101
},
{
"epoch": 51.0,
"learning_rate": 3.0263157894736844e-05,
"loss": 0.0615,
"step": 102
},
{
"epoch": 51.5,
"learning_rate": 3.0043859649122808e-05,
"loss": 0.0564,
"step": 103
},
{
"epoch": 52.0,
"learning_rate": 2.9824561403508772e-05,
"loss": 0.0621,
"step": 104
},
{
"epoch": 52.5,
"learning_rate": 2.9605263157894735e-05,
"loss": 0.0586,
"step": 105
},
{
"epoch": 53.0,
"learning_rate": 2.9385964912280706e-05,
"loss": 0.0559,
"step": 106
},
{
"epoch": 53.5,
"learning_rate": 2.916666666666667e-05,
"loss": 0.0545,
"step": 107
},
{
"epoch": 54.0,
"learning_rate": 2.8947368421052634e-05,
"loss": 0.0576,
"step": 108
},
{
"epoch": 54.5,
"learning_rate": 2.8728070175438597e-05,
"loss": 0.0547,
"step": 109
},
{
"epoch": 55.0,
"learning_rate": 2.850877192982456e-05,
"loss": 0.0575,
"step": 110
},
{
"epoch": 55.0,
"eval_accuracy_dropoff": 0.3125713892520471,
"eval_accuracy_undropoff": 0.993698363246124,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.2729271809661139,
"eval_iou_undropoff": 0.9649217226523994,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.13126148283481598,
"eval_mean_accuracy": 0.6531348762490856,
"eval_mean_iou": 0.6189244518092567,
"eval_overall_accuracy": 0.9653776168823243,
"eval_runtime": 6.3082,
"eval_samples_per_second": 3.17,
"eval_steps_per_second": 0.317,
"step": 110
},
{
"epoch": 55.5,
"learning_rate": 2.8289473684210528e-05,
"loss": 0.0548,
"step": 111
},
{
"epoch": 56.0,
"learning_rate": 2.8070175438596492e-05,
"loss": 0.0551,
"step": 112
},
{
"epoch": 56.5,
"learning_rate": 2.7850877192982456e-05,
"loss": 0.0593,
"step": 113
},
{
"epoch": 57.0,
"learning_rate": 2.7631578947368426e-05,
"loss": 0.0489,
"step": 114
},
{
"epoch": 57.5,
"learning_rate": 2.741228070175439e-05,
"loss": 0.0536,
"step": 115
},
{
"epoch": 58.0,
"learning_rate": 2.7192982456140354e-05,
"loss": 0.0562,
"step": 116
},
{
"epoch": 58.5,
"learning_rate": 2.6973684210526317e-05,
"loss": 0.0533,
"step": 117
},
{
"epoch": 59.0,
"learning_rate": 2.675438596491228e-05,
"loss": 0.0494,
"step": 118
},
{
"epoch": 59.5,
"learning_rate": 2.6535087719298245e-05,
"loss": 0.0518,
"step": 119
},
{
"epoch": 60.0,
"learning_rate": 2.6315789473684212e-05,
"loss": 0.0527,
"step": 120
},
{
"epoch": 60.0,
"eval_accuracy_dropoff": 0.3390628225418014,
"eval_accuracy_undropoff": 0.9919562338242567,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.28602939469223804,
"eval_iou_undropoff": 0.9643062074697338,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.12983748316764832,
"eval_mean_accuracy": 0.665509528183029,
"eval_mean_iou": 0.6251678010809859,
"eval_overall_accuracy": 0.9648094177246094,
"eval_runtime": 13.8883,
"eval_samples_per_second": 1.44,
"eval_steps_per_second": 0.144,
"step": 120
},
{
"epoch": 60.5,
"learning_rate": 2.6096491228070176e-05,
"loss": 0.0487,
"step": 121
},
{
"epoch": 61.0,
"learning_rate": 2.5877192982456143e-05,
"loss": 0.0562,
"step": 122
},
{
"epoch": 61.5,
"learning_rate": 2.565789473684211e-05,
"loss": 0.0496,
"step": 123
},
{
"epoch": 62.0,
"learning_rate": 2.5438596491228074e-05,
"loss": 0.0513,
"step": 124
},
{
"epoch": 62.5,
"learning_rate": 2.5219298245614037e-05,
"loss": 0.0518,
"step": 125
},
{
"epoch": 63.0,
"learning_rate": 2.5e-05,
"loss": 0.0512,
"step": 126
},
{
"epoch": 63.5,
"learning_rate": 2.4780701754385965e-05,
"loss": 0.0454,
"step": 127
},
{
"epoch": 64.0,
"learning_rate": 2.456140350877193e-05,
"loss": 0.0523,
"step": 128
},
{
"epoch": 64.5,
"learning_rate": 2.4342105263157896e-05,
"loss": 0.0483,
"step": 129
},
{
"epoch": 65.0,
"learning_rate": 2.412280701754386e-05,
"loss": 0.0491,
"step": 130
},
{
"epoch": 65.0,
"eval_accuracy_dropoff": 0.3063281267918989,
"eval_accuracy_undropoff": 0.9920477782078595,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.2588756134814734,
"eval_iou_undropoff": 0.963065642833159,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.13133761286735535,
"eval_mean_accuracy": 0.6491879524998792,
"eval_mean_iou": 0.6109706281573162,
"eval_overall_accuracy": 0.9635360717773438,
"eval_runtime": 2.8172,
"eval_samples_per_second": 7.099,
"eval_steps_per_second": 0.71,
"step": 130
},
{
"epoch": 65.5,
"learning_rate": 2.3903508771929827e-05,
"loss": 0.0477,
"step": 131
},
{
"epoch": 66.0,
"learning_rate": 2.368421052631579e-05,
"loss": 0.0476,
"step": 132
},
{
"epoch": 66.5,
"learning_rate": 2.3464912280701758e-05,
"loss": 0.0524,
"step": 133
},
{
"epoch": 67.0,
"learning_rate": 2.324561403508772e-05,
"loss": 0.0453,
"step": 134
},
{
"epoch": 67.5,
"learning_rate": 2.3026315789473685e-05,
"loss": 0.0456,
"step": 135
},
{
"epoch": 68.0,
"learning_rate": 2.280701754385965e-05,
"loss": 0.0508,
"step": 136
},
{
"epoch": 68.5,
"learning_rate": 2.2587719298245616e-05,
"loss": 0.0434,
"step": 137
},
{
"epoch": 69.0,
"learning_rate": 2.236842105263158e-05,
"loss": 0.0484,
"step": 138
},
{
"epoch": 69.5,
"learning_rate": 2.2149122807017543e-05,
"loss": 0.0459,
"step": 139
},
{
"epoch": 70.0,
"learning_rate": 2.1929824561403507e-05,
"loss": 0.0441,
"step": 140
},
{
"epoch": 70.0,
"eval_accuracy_dropoff": 0.2919333012225051,
"eval_accuracy_undropoff": 0.9939467271390291,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.2561873017406566,
"eval_iou_undropoff": 0.9643244954858242,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.12954853475093842,
"eval_mean_accuracy": 0.6429400141807671,
"eval_mean_iou": 0.6102558986132405,
"eval_overall_accuracy": 0.9647575378417969,
"eval_runtime": 2.8094,
"eval_samples_per_second": 7.119,
"eval_steps_per_second": 0.712,
"step": 140
},
{
"epoch": 70.5,
"learning_rate": 2.1710526315789474e-05,
"loss": 0.0445,
"step": 141
},
{
"epoch": 71.0,
"learning_rate": 2.149122807017544e-05,
"loss": 0.0447,
"step": 142
},
{
"epoch": 71.5,
"learning_rate": 2.1271929824561405e-05,
"loss": 0.0415,
"step": 143
},
{
"epoch": 72.0,
"learning_rate": 2.105263157894737e-05,
"loss": 0.0451,
"step": 144
},
{
"epoch": 72.5,
"learning_rate": 2.0833333333333336e-05,
"loss": 0.0433,
"step": 145
},
{
"epoch": 73.0,
"learning_rate": 2.06140350877193e-05,
"loss": 0.0411,
"step": 146
},
{
"epoch": 73.5,
"learning_rate": 2.0394736842105264e-05,
"loss": 0.0418,
"step": 147
},
{
"epoch": 74.0,
"learning_rate": 2.0175438596491227e-05,
"loss": 0.0477,
"step": 148
},
{
"epoch": 74.5,
"learning_rate": 1.9956140350877194e-05,
"loss": 0.0404,
"step": 149
},
{
"epoch": 75.0,
"learning_rate": 1.9736842105263158e-05,
"loss": 0.0426,
"step": 150
},
{
"epoch": 75.0,
"eval_accuracy_dropoff": 0.3333470951168605,
"eval_accuracy_undropoff": 0.9932967620154491,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.2887340172760432,
"eval_iou_undropoff": 0.9653766615372832,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.12331400066614151,
"eval_mean_accuracy": 0.6633219285661548,
"eval_mean_iou": 0.6270553394066631,
"eval_overall_accuracy": 0.9658565521240234,
"eval_runtime": 6.1041,
"eval_samples_per_second": 3.276,
"eval_steps_per_second": 0.328,
"step": 150
},
{
"epoch": 75.5,
"learning_rate": 1.9517543859649122e-05,
"loss": 0.0425,
"step": 151
},
{
"epoch": 76.0,
"learning_rate": 1.929824561403509e-05,
"loss": 0.0431,
"step": 152
},
{
"epoch": 76.5,
"learning_rate": 1.9078947368421056e-05,
"loss": 0.0389,
"step": 153
},
{
"epoch": 77.0,
"learning_rate": 1.885964912280702e-05,
"loss": 0.0433,
"step": 154
},
{
"epoch": 77.5,
"learning_rate": 1.8640350877192984e-05,
"loss": 0.043,
"step": 155
},
{
"epoch": 78.0,
"learning_rate": 1.8421052631578947e-05,
"loss": 0.0383,
"step": 156
},
{
"epoch": 78.5,
"learning_rate": 1.8201754385964914e-05,
"loss": 0.0412,
"step": 157
},
{
"epoch": 79.0,
"learning_rate": 1.7982456140350878e-05,
"loss": 0.0412,
"step": 158
},
{
"epoch": 79.5,
"learning_rate": 1.7763157894736842e-05,
"loss": 0.0402,
"step": 159
},
{
"epoch": 80.0,
"learning_rate": 1.7543859649122806e-05,
"loss": 0.0477,
"step": 160
},
{
"epoch": 80.0,
"eval_accuracy_dropoff": 0.33281956008165325,
"eval_accuracy_undropoff": 0.9929140666900834,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.28609114317372564,
"eval_iou_undropoff": 0.9649832592793454,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.1285596638917923,
"eval_mean_accuracy": 0.6628668133858684,
"eval_mean_iou": 0.6255372012265354,
"eval_overall_accuracy": 0.9654678344726563,
"eval_runtime": 8.3549,
"eval_samples_per_second": 2.394,
"eval_steps_per_second": 0.239,
"step": 160
},
{
"epoch": 80.5,
"learning_rate": 1.7324561403508773e-05,
"loss": 0.0409,
"step": 161
},
{
"epoch": 81.0,
"learning_rate": 1.7105263157894737e-05,
"loss": 0.0416,
"step": 162
},
{
"epoch": 81.5,
"learning_rate": 1.6885964912280704e-05,
"loss": 0.0381,
"step": 163
},
{
"epoch": 82.0,
"learning_rate": 1.6666666666666667e-05,
"loss": 0.0392,
"step": 164
},
{
"epoch": 82.5,
"learning_rate": 1.6447368421052635e-05,
"loss": 0.0398,
"step": 165
},
{
"epoch": 83.0,
"learning_rate": 1.62280701754386e-05,
"loss": 0.0416,
"step": 166
},
{
"epoch": 83.5,
"learning_rate": 1.6008771929824562e-05,
"loss": 0.0389,
"step": 167
},
{
"epoch": 84.0,
"learning_rate": 1.5789473684210526e-05,
"loss": 0.0398,
"step": 168
},
{
"epoch": 84.5,
"learning_rate": 1.5570175438596493e-05,
"loss": 0.0402,
"step": 169
},
{
"epoch": 85.0,
"learning_rate": 1.5350877192982457e-05,
"loss": 0.039,
"step": 170
},
{
"epoch": 85.0,
"eval_accuracy_dropoff": 0.37348104314319136,
"eval_accuracy_undropoff": 0.9912656707566442,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.31088955415374747,
"eval_iou_undropoff": 0.9650356885247457,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.12654618918895721,
"eval_mean_accuracy": 0.6823733569499177,
"eval_mean_iou": 0.6379626213392466,
"eval_overall_accuracy": 0.9655786514282226,
"eval_runtime": 2.7731,
"eval_samples_per_second": 7.212,
"eval_steps_per_second": 0.721,
"step": 170
},
{
"epoch": 85.5,
"learning_rate": 1.5131578947368422e-05,
"loss": 0.0394,
"step": 171
},
{
"epoch": 86.0,
"learning_rate": 1.4912280701754386e-05,
"loss": 0.0377,
"step": 172
},
{
"epoch": 86.5,
"learning_rate": 1.4692982456140353e-05,
"loss": 0.0393,
"step": 173
},
{
"epoch": 87.0,
"learning_rate": 1.4473684210526317e-05,
"loss": 0.0379,
"step": 174
},
{
"epoch": 87.5,
"learning_rate": 1.425438596491228e-05,
"loss": 0.0387,
"step": 175
},
{
"epoch": 88.0,
"learning_rate": 1.4035087719298246e-05,
"loss": 0.0388,
"step": 176
},
{
"epoch": 88.5,
"learning_rate": 1.3815789473684213e-05,
"loss": 0.0381,
"step": 177
},
{
"epoch": 89.0,
"learning_rate": 1.3596491228070177e-05,
"loss": 0.0362,
"step": 178
},
{
"epoch": 89.5,
"learning_rate": 1.337719298245614e-05,
"loss": 0.0378,
"step": 179
},
{
"epoch": 90.0,
"learning_rate": 1.3157894736842106e-05,
"loss": 0.0378,
"step": 180
},
{
"epoch": 90.0,
"eval_accuracy_dropoff": 0.3154154911809904,
"eval_accuracy_undropoff": 0.9931711870022896,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.2725190143831572,
"eval_iou_undropoff": 0.9645253754212552,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.13092683255672455,
"eval_mean_accuracy": 0.65429333909164,
"eval_mean_iou": 0.6185221949022062,
"eval_overall_accuracy": 0.9649906158447266,
"eval_runtime": 2.7934,
"eval_samples_per_second": 7.16,
"eval_steps_per_second": 0.716,
"step": 180
},
{
"epoch": 90.5,
"learning_rate": 1.2938596491228071e-05,
"loss": 0.0369,
"step": 181
},
{
"epoch": 91.0,
"learning_rate": 1.2719298245614037e-05,
"loss": 0.0387,
"step": 182
},
{
"epoch": 91.5,
"learning_rate": 1.25e-05,
"loss": 0.0371,
"step": 183
},
{
"epoch": 92.0,
"learning_rate": 1.2280701754385964e-05,
"loss": 0.0356,
"step": 184
},
{
"epoch": 92.5,
"learning_rate": 1.206140350877193e-05,
"loss": 0.0377,
"step": 185
},
{
"epoch": 93.0,
"learning_rate": 1.1842105263157895e-05,
"loss": 0.0337,
"step": 186
},
{
"epoch": 93.5,
"learning_rate": 1.162280701754386e-05,
"loss": 0.038,
"step": 187
},
{
"epoch": 94.0,
"learning_rate": 1.1403508771929824e-05,
"loss": 0.0346,
"step": 188
},
{
"epoch": 94.5,
"learning_rate": 1.118421052631579e-05,
"loss": 0.036,
"step": 189
},
{
"epoch": 95.0,
"learning_rate": 1.0964912280701754e-05,
"loss": 0.0362,
"step": 190
},
{
"epoch": 95.0,
"eval_accuracy_dropoff": 0.3507511640175233,
"eval_accuracy_undropoff": 0.992195642288331,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.297273444759361,
"eval_iou_undropoff": 0.9650146387690659,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.12657637894153595,
"eval_mean_accuracy": 0.6714734031529271,
"eval_mean_iou": 0.6311440417642135,
"eval_overall_accuracy": 0.9655248641967773,
"eval_runtime": 4.6147,
"eval_samples_per_second": 4.334,
"eval_steps_per_second": 0.433,
"step": 190
},
{
"epoch": 95.5,
"learning_rate": 1.074561403508772e-05,
"loss": 0.0373,
"step": 191
},
{
"epoch": 96.0,
"learning_rate": 1.0526315789473684e-05,
"loss": 0.0358,
"step": 192
},
{
"epoch": 96.5,
"learning_rate": 1.030701754385965e-05,
"loss": 0.0372,
"step": 193
},
{
"epoch": 97.0,
"learning_rate": 1.0087719298245614e-05,
"loss": 0.0371,
"step": 194
},
{
"epoch": 97.5,
"learning_rate": 9.868421052631579e-06,
"loss": 0.0363,
"step": 195
},
{
"epoch": 98.0,
"learning_rate": 9.649122807017545e-06,
"loss": 0.0376,
"step": 196
},
{
"epoch": 98.5,
"learning_rate": 9.42982456140351e-06,
"loss": 0.0371,
"step": 197
},
{
"epoch": 99.0,
"learning_rate": 9.210526315789474e-06,
"loss": 0.0427,
"step": 198
},
{
"epoch": 99.5,
"learning_rate": 8.991228070175439e-06,
"loss": 0.0334,
"step": 199
},
{
"epoch": 100.0,
"learning_rate": 8.771929824561403e-06,
"loss": 0.0394,
"step": 200
},
{
"epoch": 100.0,
"eval_accuracy_dropoff": 0.33373701231679626,
"eval_accuracy_undropoff": 0.993359051998205,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.28943170290215425,
"eval_iou_undropoff": 0.9654530730096186,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.1306542307138443,
"eval_mean_accuracy": 0.6635480321575006,
"eval_mean_iou": 0.6274423879558864,
"eval_overall_accuracy": 0.9659324645996094,
"eval_runtime": 11.5796,
"eval_samples_per_second": 1.727,
"eval_steps_per_second": 0.173,
"step": 200
},
{
"epoch": 100.5,
"learning_rate": 8.552631578947368e-06,
"loss": 0.0349,
"step": 201
},
{
"epoch": 101.0,
"learning_rate": 8.333333333333334e-06,
"loss": 0.037,
"step": 202
},
{
"epoch": 101.5,
"learning_rate": 8.1140350877193e-06,
"loss": 0.0345,
"step": 203
},
{
"epoch": 102.0,
"learning_rate": 7.894736842105263e-06,
"loss": 0.0371,
"step": 204
},
{
"epoch": 102.5,
"learning_rate": 7.675438596491228e-06,
"loss": 0.0355,
"step": 205
},
{
"epoch": 103.0,
"learning_rate": 7.456140350877193e-06,
"loss": 0.0346,
"step": 206
},
{
"epoch": 103.5,
"learning_rate": 7.236842105263158e-06,
"loss": 0.0363,
"step": 207
},
{
"epoch": 104.0,
"learning_rate": 7.017543859649123e-06,
"loss": 0.0348,
"step": 208
},
{
"epoch": 104.5,
"learning_rate": 6.798245614035088e-06,
"loss": 0.0351,
"step": 209
},
{
"epoch": 105.0,
"learning_rate": 6.578947368421053e-06,
"loss": 0.0362,
"step": 210
},
{
"epoch": 105.0,
"eval_accuracy_dropoff": 0.3660680290832359,
"eval_accuracy_undropoff": 0.9918169271535567,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.3079764119267967,
"eval_iou_undropoff": 0.965270142099317,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.12709322571754456,
"eval_mean_accuracy": 0.6789424781183964,
"eval_mean_iou": 0.6366232770130569,
"eval_overall_accuracy": 0.9657987594604492,
"eval_runtime": 2.7866,
"eval_samples_per_second": 7.177,
"eval_steps_per_second": 0.718,
"step": 210
},
{
"epoch": 105.5,
"learning_rate": 6.3596491228070184e-06,
"loss": 0.0347,
"step": 211
},
{
"epoch": 106.0,
"learning_rate": 6.140350877192982e-06,
"loss": 0.036,
"step": 212
},
{
"epoch": 106.5,
"learning_rate": 5.921052631578948e-06,
"loss": 0.0355,
"step": 213
},
{
"epoch": 107.0,
"learning_rate": 5.701754385964912e-06,
"loss": 0.0345,
"step": 214
},
{
"epoch": 107.5,
"learning_rate": 5.482456140350877e-06,
"loss": 0.0334,
"step": 215
},
{
"epoch": 108.0,
"learning_rate": 5.263157894736842e-06,
"loss": 0.0391,
"step": 216
},
{
"epoch": 108.5,
"learning_rate": 5.043859649122807e-06,
"loss": 0.0352,
"step": 217
},
{
"epoch": 109.0,
"learning_rate": 4.824561403508772e-06,
"loss": 0.0343,
"step": 218
},
{
"epoch": 109.5,
"learning_rate": 4.605263157894737e-06,
"loss": 0.0331,
"step": 219
},
{
"epoch": 110.0,
"learning_rate": 4.3859649122807014e-06,
"loss": 0.0361,
"step": 220
},
{
"epoch": 110.0,
"eval_accuracy_dropoff": 0.35535677423794126,
"eval_accuracy_undropoff": 0.9917641896282203,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.29865949055243485,
"eval_iou_undropoff": 0.9647824937168638,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.12739057838916779,
"eval_mean_accuracy": 0.6735604819330808,
"eval_mean_iou": 0.6317209921346493,
"eval_overall_accuracy": 0.965302848815918,
"eval_runtime": 2.8789,
"eval_samples_per_second": 6.947,
"eval_steps_per_second": 0.695,
"step": 220
},
{
"epoch": 110.5,
"learning_rate": 4.166666666666667e-06,
"loss": 0.0354,
"step": 221
},
{
"epoch": 111.0,
"learning_rate": 3.9473684210526315e-06,
"loss": 0.0334,
"step": 222
},
{
"epoch": 111.5,
"learning_rate": 3.7280701754385965e-06,
"loss": 0.0333,
"step": 223
},
{
"epoch": 112.0,
"learning_rate": 3.5087719298245615e-06,
"loss": 0.0401,
"step": 224
},
{
"epoch": 112.5,
"learning_rate": 3.2894736842105265e-06,
"loss": 0.0313,
"step": 225
},
{
"epoch": 113.0,
"learning_rate": 3.070175438596491e-06,
"loss": 0.0416,
"step": 226
},
{
"epoch": 113.5,
"learning_rate": 2.850877192982456e-06,
"loss": 0.0349,
"step": 227
},
{
"epoch": 114.0,
"learning_rate": 2.631578947368421e-06,
"loss": 0.0357,
"step": 228
},
{
"epoch": 114.5,
"learning_rate": 2.412280701754386e-06,
"loss": 0.0329,
"step": 229
},
{
"epoch": 115.0,
"learning_rate": 2.1929824561403507e-06,
"loss": 0.0353,
"step": 230
},
{
"epoch": 115.0,
"eval_accuracy_dropoff": 0.32280556893506734,
"eval_accuracy_undropoff": 0.9930850158759852,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.2784262212058146,
"eval_iou_undropoff": 0.9647420691308475,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.12896691262722015,
"eval_mean_accuracy": 0.6579452924055262,
"eval_mean_iou": 0.621584145168331,
"eval_overall_accuracy": 0.9652153015136719,
"eval_runtime": 5.6606,
"eval_samples_per_second": 3.533,
"eval_steps_per_second": 0.353,
"step": 230
},
{
"epoch": 115.5,
"learning_rate": 1.9736842105263157e-06,
"loss": 0.0334,
"step": 231
},
{
"epoch": 116.0,
"learning_rate": 1.7543859649122807e-06,
"loss": 0.0372,
"step": 232
},
{
"epoch": 116.5,
"learning_rate": 1.5350877192982455e-06,
"loss": 0.0356,
"step": 233
},
{
"epoch": 117.0,
"learning_rate": 1.3157894736842106e-06,
"loss": 0.033,
"step": 234
},
{
"epoch": 117.5,
"learning_rate": 1.0964912280701754e-06,
"loss": 0.0339,
"step": 235
},
{
"epoch": 118.0,
"learning_rate": 8.771929824561404e-07,
"loss": 0.0341,
"step": 236
},
{
"epoch": 118.5,
"learning_rate": 6.578947368421053e-07,
"loss": 0.0386,
"step": 237
},
{
"epoch": 119.0,
"learning_rate": 4.385964912280702e-07,
"loss": 0.034,
"step": 238
},
{
"epoch": 119.5,
"learning_rate": 2.192982456140351e-07,
"loss": 0.0335,
"step": 239
},
{
"epoch": 120.0,
"learning_rate": 0.0,
"loss": 0.0344,
"step": 240
},
{
"epoch": 120.0,
"eval_accuracy_dropoff": 0.33194339319709165,
"eval_accuracy_undropoff": 0.9926412246250411,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.28380371178011704,
"eval_iou_undropoff": 0.9646824554135581,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.12955032289028168,
"eval_mean_accuracy": 0.6622923089110664,
"eval_mean_iou": 0.6242430835968376,
"eval_overall_accuracy": 0.9651699066162109,
"eval_runtime": 11.3006,
"eval_samples_per_second": 1.77,
"eval_steps_per_second": 0.177,
"step": 240
}
],
"max_steps": 240,
"num_train_epochs": 120,
"total_flos": 2.778548611716219e+18,
"trial_name": null,
"trial_params": null
}