sam1120's picture
Training in progress, step 120
20596da
raw
history blame
22.1 kB
{
"best_metric": 0.7164114117622375,
"best_model_checkpoint": "/robodata/smodak/Projects/nspl/scripts/terrainseg/training/models/dropoff-utcustom-train-SF-RGB-b5_1/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": 1.6666666666666665e-07,
"loss": 1.0244,
"step": 1
},
{
"epoch": 1.0,
"learning_rate": 3.333333333333333e-07,
"loss": 1.0185,
"step": 2
},
{
"epoch": 1.5,
"learning_rate": 5e-07,
"loss": 1.0196,
"step": 3
},
{
"epoch": 2.0,
"learning_rate": 6.666666666666666e-07,
"loss": 1.0239,
"step": 4
},
{
"epoch": 2.5,
"learning_rate": 8.333333333333333e-07,
"loss": 1.0206,
"step": 5
},
{
"epoch": 3.0,
"learning_rate": 1e-06,
"loss": 1.0123,
"step": 6
},
{
"epoch": 3.5,
"learning_rate": 1.1666666666666668e-06,
"loss": 1.0125,
"step": 7
},
{
"epoch": 4.0,
"learning_rate": 1.3333333333333332e-06,
"loss": 1.016,
"step": 8
},
{
"epoch": 4.5,
"learning_rate": 1.5e-06,
"loss": 1.007,
"step": 9
},
{
"epoch": 5.0,
"learning_rate": 1.6666666666666667e-06,
"loss": 1.0071,
"step": 10
},
{
"epoch": 5.0,
"eval_accuracy_dropoff": 0.025546457487557055,
"eval_accuracy_undropoff": 0.5241011485834999,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.014703669184605108,
"eval_iou_undropoff": 0.5086504735858094,
"eval_iou_unlabeled": 0.0,
"eval_loss": 1.0205577611923218,
"eval_mean_accuracy": 0.27482380303552845,
"eval_mean_iou": 0.17445138092347148,
"eval_overall_accuracy": 0.5033716201782227,
"eval_runtime": 2.8831,
"eval_samples_per_second": 6.937,
"eval_steps_per_second": 0.694,
"step": 10
},
{
"epoch": 5.5,
"learning_rate": 1.833333333333333e-06,
"loss": 1.002,
"step": 11
},
{
"epoch": 6.0,
"learning_rate": 2e-06,
"loss": 0.9999,
"step": 12
},
{
"epoch": 6.5,
"learning_rate": 1.9912280701754384e-06,
"loss": 0.9963,
"step": 13
},
{
"epoch": 7.0,
"learning_rate": 1.982456140350877e-06,
"loss": 0.9889,
"step": 14
},
{
"epoch": 7.5,
"learning_rate": 1.9736842105263157e-06,
"loss": 0.9863,
"step": 15
},
{
"epoch": 8.0,
"learning_rate": 1.964912280701754e-06,
"loss": 0.9881,
"step": 16
},
{
"epoch": 8.5,
"learning_rate": 1.956140350877193e-06,
"loss": 0.9856,
"step": 17
},
{
"epoch": 9.0,
"learning_rate": 1.9473684210526315e-06,
"loss": 0.9763,
"step": 18
},
{
"epoch": 9.5,
"learning_rate": 1.93859649122807e-06,
"loss": 0.9789,
"step": 19
},
{
"epoch": 10.0,
"learning_rate": 1.929824561403509e-06,
"loss": 0.9688,
"step": 20
},
{
"epoch": 10.0,
"eval_accuracy_dropoff": 0.09924998279777059,
"eval_accuracy_undropoff": 0.5978783594052401,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.05823063486894339,
"eval_iou_undropoff": 0.5837983317382666,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.9873144030570984,
"eval_mean_accuracy": 0.3485641711015054,
"eval_mean_iou": 0.21400965553573667,
"eval_overall_accuracy": 0.577145767211914,
"eval_runtime": 2.4113,
"eval_samples_per_second": 8.294,
"eval_steps_per_second": 0.829,
"step": 20
},
{
"epoch": 10.5,
"learning_rate": 1.9210526315789474e-06,
"loss": 0.9671,
"step": 21
},
{
"epoch": 11.0,
"learning_rate": 1.912280701754386e-06,
"loss": 0.9643,
"step": 22
},
{
"epoch": 11.5,
"learning_rate": 1.9035087719298243e-06,
"loss": 0.9621,
"step": 23
},
{
"epoch": 12.0,
"learning_rate": 1.894736842105263e-06,
"loss": 0.9631,
"step": 24
},
{
"epoch": 12.5,
"learning_rate": 1.8859649122807019e-06,
"loss": 0.958,
"step": 25
},
{
"epoch": 13.0,
"learning_rate": 1.8771929824561403e-06,
"loss": 0.9523,
"step": 26
},
{
"epoch": 13.5,
"learning_rate": 1.868421052631579e-06,
"loss": 0.9526,
"step": 27
},
{
"epoch": 14.0,
"learning_rate": 1.8596491228070175e-06,
"loss": 0.9428,
"step": 28
},
{
"epoch": 14.5,
"learning_rate": 1.850877192982456e-06,
"loss": 0.947,
"step": 29
},
{
"epoch": 15.0,
"learning_rate": 1.8421052631578946e-06,
"loss": 0.9406,
"step": 30
},
{
"epoch": 15.0,
"eval_accuracy_dropoff": 0.2037844904699649,
"eval_accuracy_undropoff": 0.6855052404184374,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.1135240392724076,
"eval_iou_undropoff": 0.6704558637017982,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.9312955141067505,
"eval_mean_accuracy": 0.4446448654442012,
"eval_mean_iou": 0.2613266343247353,
"eval_overall_accuracy": 0.6654756546020508,
"eval_runtime": 2.3961,
"eval_samples_per_second": 8.347,
"eval_steps_per_second": 0.835,
"step": 30
},
{
"epoch": 15.5,
"learning_rate": 1.833333333333333e-06,
"loss": 0.9434,
"step": 31
},
{
"epoch": 16.0,
"learning_rate": 1.8245614035087718e-06,
"loss": 0.9337,
"step": 32
},
{
"epoch": 16.5,
"learning_rate": 1.8157894736842106e-06,
"loss": 0.9365,
"step": 33
},
{
"epoch": 17.0,
"learning_rate": 1.807017543859649e-06,
"loss": 0.9311,
"step": 34
},
{
"epoch": 17.5,
"learning_rate": 1.7982456140350878e-06,
"loss": 0.9313,
"step": 35
},
{
"epoch": 18.0,
"learning_rate": 1.7894736842105262e-06,
"loss": 0.9281,
"step": 36
},
{
"epoch": 18.5,
"learning_rate": 1.7807017543859647e-06,
"loss": 0.927,
"step": 37
},
{
"epoch": 19.0,
"learning_rate": 1.7719298245614034e-06,
"loss": 0.9209,
"step": 38
},
{
"epoch": 19.5,
"learning_rate": 1.7631578947368419e-06,
"loss": 0.9161,
"step": 39
},
{
"epoch": 20.0,
"learning_rate": 1.7543859649122805e-06,
"loss": 0.9278,
"step": 40
},
{
"epoch": 20.0,
"eval_accuracy_dropoff": 0.30091515860455514,
"eval_accuracy_undropoff": 0.7288984723033463,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.16481371211493093,
"eval_iou_undropoff": 0.7141641818866753,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.88507080078125,
"eval_mean_accuracy": 0.5149068154539507,
"eval_mean_iou": 0.2929926313338687,
"eval_overall_accuracy": 0.7111032485961915,
"eval_runtime": 2.3253,
"eval_samples_per_second": 8.601,
"eval_steps_per_second": 0.86,
"step": 40
},
{
"epoch": 20.5,
"learning_rate": 1.7456140350877194e-06,
"loss": 0.9123,
"step": 41
},
{
"epoch": 21.0,
"learning_rate": 1.7368421052631579e-06,
"loss": 0.9098,
"step": 42
},
{
"epoch": 21.5,
"learning_rate": 1.7280701754385966e-06,
"loss": 0.9114,
"step": 43
},
{
"epoch": 22.0,
"learning_rate": 1.719298245614035e-06,
"loss": 0.9087,
"step": 44
},
{
"epoch": 22.5,
"learning_rate": 1.7105263157894735e-06,
"loss": 0.9003,
"step": 45
},
{
"epoch": 23.0,
"learning_rate": 1.7017543859649122e-06,
"loss": 0.9087,
"step": 46
},
{
"epoch": 23.5,
"learning_rate": 1.6929824561403506e-06,
"loss": 0.901,
"step": 47
},
{
"epoch": 24.0,
"learning_rate": 1.6842105263157893e-06,
"loss": 0.8953,
"step": 48
},
{
"epoch": 24.5,
"learning_rate": 1.6754385964912282e-06,
"loss": 0.8924,
"step": 49
},
{
"epoch": 25.0,
"learning_rate": 1.6666666666666667e-06,
"loss": 0.8956,
"step": 50
},
{
"epoch": 25.0,
"eval_accuracy_dropoff": 0.3769627743755591,
"eval_accuracy_undropoff": 0.7513628670108868,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.19850953583032382,
"eval_iou_undropoff": 0.7370141563659725,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.8563140630722046,
"eval_mean_accuracy": 0.564162820693223,
"eval_mean_iou": 0.31184123073209874,
"eval_overall_accuracy": 0.7357955932617187,
"eval_runtime": 2.2641,
"eval_samples_per_second": 8.834,
"eval_steps_per_second": 0.883,
"step": 50
},
{
"epoch": 25.5,
"learning_rate": 1.6578947368421053e-06,
"loss": 0.8909,
"step": 51
},
{
"epoch": 26.0,
"learning_rate": 1.6491228070175438e-06,
"loss": 0.8821,
"step": 52
},
{
"epoch": 26.5,
"learning_rate": 1.6403508771929823e-06,
"loss": 0.8869,
"step": 53
},
{
"epoch": 27.0,
"learning_rate": 1.631578947368421e-06,
"loss": 0.8838,
"step": 54
},
{
"epoch": 27.5,
"learning_rate": 1.6228070175438594e-06,
"loss": 0.8769,
"step": 55
},
{
"epoch": 28.0,
"learning_rate": 1.614035087719298e-06,
"loss": 0.8819,
"step": 56
},
{
"epoch": 28.5,
"learning_rate": 1.605263157894737e-06,
"loss": 0.8742,
"step": 57
},
{
"epoch": 29.0,
"learning_rate": 1.5964912280701754e-06,
"loss": 0.8858,
"step": 58
},
{
"epoch": 29.5,
"learning_rate": 1.5877192982456141e-06,
"loss": 0.8734,
"step": 59
},
{
"epoch": 30.0,
"learning_rate": 1.5789473684210526e-06,
"loss": 0.8674,
"step": 60
},
{
"epoch": 30.0,
"eval_accuracy_dropoff": 0.43656047157044886,
"eval_accuracy_undropoff": 0.7806618459924953,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.2245523958736421,
"eval_iou_undropoff": 0.7664355612520527,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.8260032534599304,
"eval_mean_accuracy": 0.6086111587814721,
"eval_mean_iou": 0.3303293190418983,
"eval_overall_accuracy": 0.7663543701171875,
"eval_runtime": 2.6403,
"eval_samples_per_second": 7.575,
"eval_steps_per_second": 0.757,
"step": 60
},
{
"epoch": 30.5,
"learning_rate": 1.570175438596491e-06,
"loss": 0.8687,
"step": 61
},
{
"epoch": 31.0,
"learning_rate": 1.5614035087719297e-06,
"loss": 0.8671,
"step": 62
},
{
"epoch": 31.5,
"learning_rate": 1.5526315789473682e-06,
"loss": 0.866,
"step": 63
},
{
"epoch": 32.0,
"learning_rate": 1.5438596491228069e-06,
"loss": 0.853,
"step": 64
},
{
"epoch": 32.5,
"learning_rate": 1.5350877192982458e-06,
"loss": 0.8616,
"step": 65
},
{
"epoch": 33.0,
"learning_rate": 1.5263157894736842e-06,
"loss": 0.8555,
"step": 66
},
{
"epoch": 33.5,
"learning_rate": 1.517543859649123e-06,
"loss": 0.856,
"step": 67
},
{
"epoch": 34.0,
"learning_rate": 1.5087719298245614e-06,
"loss": 0.8452,
"step": 68
},
{
"epoch": 34.5,
"learning_rate": 1.5e-06,
"loss": 0.8559,
"step": 69
},
{
"epoch": 35.0,
"learning_rate": 1.4912280701754385e-06,
"loss": 0.8438,
"step": 70
},
{
"epoch": 35.0,
"eval_accuracy_dropoff": 0.4920571572742494,
"eval_accuracy_undropoff": 0.7790013104777522,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.23805307181361615,
"eval_iou_undropoff": 0.7660370749081444,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.8148576617240906,
"eval_mean_accuracy": 0.6355292338760008,
"eval_mean_iou": 0.3346967155739202,
"eval_overall_accuracy": 0.7670703887939453,
"eval_runtime": 2.6339,
"eval_samples_per_second": 7.593,
"eval_steps_per_second": 0.759,
"step": 70
},
{
"epoch": 35.5,
"learning_rate": 1.482456140350877e-06,
"loss": 0.8387,
"step": 71
},
{
"epoch": 36.0,
"learning_rate": 1.4736842105263156e-06,
"loss": 0.8475,
"step": 72
},
{
"epoch": 36.5,
"learning_rate": 1.4649122807017541e-06,
"loss": 0.8298,
"step": 73
},
{
"epoch": 37.0,
"learning_rate": 1.456140350877193e-06,
"loss": 0.8521,
"step": 74
},
{
"epoch": 37.5,
"learning_rate": 1.4473684210526317e-06,
"loss": 0.8356,
"step": 75
},
{
"epoch": 38.0,
"learning_rate": 1.4385964912280701e-06,
"loss": 0.8229,
"step": 76
},
{
"epoch": 38.5,
"learning_rate": 1.4298245614035088e-06,
"loss": 0.8277,
"step": 77
},
{
"epoch": 39.0,
"learning_rate": 1.4210526315789473e-06,
"loss": 0.8268,
"step": 78
},
{
"epoch": 39.5,
"learning_rate": 1.4122807017543857e-06,
"loss": 0.8204,
"step": 79
},
{
"epoch": 40.0,
"learning_rate": 1.4035087719298244e-06,
"loss": 0.8309,
"step": 80
},
{
"epoch": 40.0,
"eval_accuracy_dropoff": 0.49719488979105025,
"eval_accuracy_undropoff": 0.797183020108918,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.25387240561121116,
"eval_iou_undropoff": 0.7839036228326106,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.7881223559379578,
"eval_mean_accuracy": 0.6471889549499841,
"eval_mean_iou": 0.34592534281460724,
"eval_overall_accuracy": 0.7847097396850586,
"eval_runtime": 2.3528,
"eval_samples_per_second": 8.501,
"eval_steps_per_second": 0.85,
"step": 80
},
{
"epoch": 40.5,
"learning_rate": 1.3947368421052629e-06,
"loss": 0.8164,
"step": 81
},
{
"epoch": 41.0,
"learning_rate": 1.3859649122807018e-06,
"loss": 0.8311,
"step": 82
},
{
"epoch": 41.5,
"learning_rate": 1.3771929824561405e-06,
"loss": 0.8189,
"step": 83
},
{
"epoch": 42.0,
"learning_rate": 1.368421052631579e-06,
"loss": 0.8179,
"step": 84
},
{
"epoch": 42.5,
"learning_rate": 1.3596491228070176e-06,
"loss": 0.8202,
"step": 85
},
{
"epoch": 43.0,
"learning_rate": 1.350877192982456e-06,
"loss": 0.8104,
"step": 86
},
{
"epoch": 43.5,
"learning_rate": 1.3421052631578945e-06,
"loss": 0.8061,
"step": 87
},
{
"epoch": 44.0,
"learning_rate": 1.3333333333333332e-06,
"loss": 0.817,
"step": 88
},
{
"epoch": 44.5,
"learning_rate": 1.3245614035087717e-06,
"loss": 0.8078,
"step": 89
},
{
"epoch": 45.0,
"learning_rate": 1.3157894736842106e-06,
"loss": 0.8069,
"step": 90
},
{
"epoch": 45.0,
"eval_accuracy_dropoff": 0.5063373013142504,
"eval_accuracy_undropoff": 0.816973124758079,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.2667705916473318,
"eval_iou_undropoff": 0.8033164328582806,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.7640160322189331,
"eval_mean_accuracy": 0.6616552130361647,
"eval_mean_iou": 0.35669567483520415,
"eval_overall_accuracy": 0.8040571212768555,
"eval_runtime": 2.4378,
"eval_samples_per_second": 8.204,
"eval_steps_per_second": 0.82,
"step": 90
},
{
"epoch": 45.5,
"learning_rate": 1.3070175438596492e-06,
"loss": 0.796,
"step": 91
},
{
"epoch": 46.0,
"learning_rate": 1.2982456140350877e-06,
"loss": 0.8078,
"step": 92
},
{
"epoch": 46.5,
"learning_rate": 1.2894736842105264e-06,
"loss": 0.8064,
"step": 93
},
{
"epoch": 47.0,
"learning_rate": 1.2807017543859648e-06,
"loss": 0.7821,
"step": 94
},
{
"epoch": 47.5,
"learning_rate": 1.2719298245614033e-06,
"loss": 0.7981,
"step": 95
},
{
"epoch": 48.0,
"learning_rate": 1.263157894736842e-06,
"loss": 0.7868,
"step": 96
},
{
"epoch": 48.5,
"learning_rate": 1.2543859649122804e-06,
"loss": 0.7887,
"step": 97
},
{
"epoch": 49.0,
"learning_rate": 1.2456140350877193e-06,
"loss": 0.7935,
"step": 98
},
{
"epoch": 49.5,
"learning_rate": 1.236842105263158e-06,
"loss": 0.7878,
"step": 99
},
{
"epoch": 50.0,
"learning_rate": 1.2280701754385965e-06,
"loss": 0.7779,
"step": 100
},
{
"epoch": 50.0,
"eval_accuracy_dropoff": 0.5315901740865616,
"eval_accuracy_undropoff": 0.8268033994807842,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.27780135060302485,
"eval_iou_undropoff": 0.8132247439558549,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.7485586404800415,
"eval_mean_accuracy": 0.6791967867836729,
"eval_mean_iou": 0.36367536485295987,
"eval_overall_accuracy": 0.8145286560058593,
"eval_runtime": 2.3681,
"eval_samples_per_second": 8.446,
"eval_steps_per_second": 0.845,
"step": 100
},
{
"epoch": 50.5,
"learning_rate": 1.2192982456140352e-06,
"loss": 0.7784,
"step": 101
},
{
"epoch": 51.0,
"learning_rate": 1.2105263157894736e-06,
"loss": 0.7842,
"step": 102
},
{
"epoch": 51.5,
"learning_rate": 1.201754385964912e-06,
"loss": 0.7646,
"step": 103
},
{
"epoch": 52.0,
"learning_rate": 1.1929824561403508e-06,
"loss": 0.7897,
"step": 104
},
{
"epoch": 52.5,
"learning_rate": 1.1842105263157892e-06,
"loss": 0.7794,
"step": 105
},
{
"epoch": 53.0,
"learning_rate": 1.1754385964912281e-06,
"loss": 0.7655,
"step": 106
},
{
"epoch": 53.5,
"learning_rate": 1.1666666666666668e-06,
"loss": 0.7677,
"step": 107
},
{
"epoch": 54.0,
"learning_rate": 1.1578947368421053e-06,
"loss": 0.7643,
"step": 108
},
{
"epoch": 54.5,
"learning_rate": 1.149122807017544e-06,
"loss": 0.7651,
"step": 109
},
{
"epoch": 55.0,
"learning_rate": 1.1403508771929824e-06,
"loss": 0.7695,
"step": 110
},
{
"epoch": 55.0,
"eval_accuracy_dropoff": 0.5541870226381339,
"eval_accuracy_undropoff": 0.8329422464394708,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.28578390091121564,
"eval_iou_undropoff": 0.8195285806313918,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.7353506684303284,
"eval_mean_accuracy": 0.6935646345388023,
"eval_mean_iou": 0.3684374938475358,
"eval_overall_accuracy": 0.8213518142700196,
"eval_runtime": 2.4084,
"eval_samples_per_second": 8.304,
"eval_steps_per_second": 0.83,
"step": 110
},
{
"epoch": 55.5,
"learning_rate": 1.1315789473684209e-06,
"loss": 0.7605,
"step": 111
},
{
"epoch": 56.0,
"learning_rate": 1.1228070175438595e-06,
"loss": 0.7552,
"step": 112
},
{
"epoch": 56.5,
"learning_rate": 1.114035087719298e-06,
"loss": 0.7735,
"step": 113
},
{
"epoch": 57.0,
"learning_rate": 1.1052631578947369e-06,
"loss": 0.7486,
"step": 114
},
{
"epoch": 57.5,
"learning_rate": 1.0964912280701756e-06,
"loss": 0.7472,
"step": 115
},
{
"epoch": 58.0,
"learning_rate": 1.087719298245614e-06,
"loss": 0.7678,
"step": 116
},
{
"epoch": 58.5,
"learning_rate": 1.0789473684210527e-06,
"loss": 0.7578,
"step": 117
},
{
"epoch": 59.0,
"learning_rate": 1.0701754385964912e-06,
"loss": 0.7411,
"step": 118
},
{
"epoch": 59.5,
"learning_rate": 1.0614035087719296e-06,
"loss": 0.751,
"step": 119
},
{
"epoch": 60.0,
"learning_rate": 1.0526315789473683e-06,
"loss": 0.7568,
"step": 120
},
{
"epoch": 60.0,
"eval_accuracy_dropoff": 0.5577375627881374,
"eval_accuracy_undropoff": 0.848569867768118,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.2924013736977288,
"eval_iou_undropoff": 0.8346954991270276,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.7164114117622375,
"eval_mean_accuracy": 0.7031537152781278,
"eval_mean_iou": 0.3756989576082521,
"eval_overall_accuracy": 0.8364772796630859,
"eval_runtime": 2.284,
"eval_samples_per_second": 8.757,
"eval_steps_per_second": 0.876,
"step": 120
}
],
"max_steps": 240,
"num_train_epochs": 120,
"total_flos": 6.945599099555021e+17,
"trial_name": null,
"trial_params": null
}