sam1120's picture
Training in progress, step 240
4300b18
raw
history blame
43.6 kB
{
"best_metric": 0.6194807887077332,
"best_model_checkpoint": "/robodata/smodak/Projects/nspl/scripts/terrainseg/training/models/dropoff-utcustom-train-SF-RGB-b0_1/checkpoint-240",
"epoch": 80.0,
"global_step": 240,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.33,
"learning_rate": 5e-07,
"loss": 1.1127,
"step": 1
},
{
"epoch": 0.67,
"learning_rate": 1e-06,
"loss": 1.1148,
"step": 2
},
{
"epoch": 1.0,
"learning_rate": 1.5e-06,
"loss": 1.109,
"step": 3
},
{
"epoch": 1.33,
"learning_rate": 2e-06,
"loss": 1.1127,
"step": 4
},
{
"epoch": 1.67,
"learning_rate": 2.5e-06,
"loss": 1.1118,
"step": 5
},
{
"epoch": 2.0,
"learning_rate": 3e-06,
"loss": 1.1088,
"step": 6
},
{
"epoch": 2.33,
"learning_rate": 3.5e-06,
"loss": 1.1122,
"step": 7
},
{
"epoch": 2.67,
"learning_rate": 4e-06,
"loss": 1.1054,
"step": 8
},
{
"epoch": 3.0,
"learning_rate": 4.5e-06,
"loss": 1.1061,
"step": 9
},
{
"epoch": 3.33,
"learning_rate": 5e-06,
"loss": 1.1029,
"step": 10
},
{
"epoch": 3.33,
"eval_accuracy_dropoff": 0.3332988719645278,
"eval_accuracy_undropoff": 0.45771106945832574,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.04096214090185886,
"eval_iou_undropoff": 0.4500892406005614,
"eval_iou_unlabeled": 0.0,
"eval_loss": 1.085178256034851,
"eval_mean_accuracy": 0.39550497071142676,
"eval_mean_iou": 0.16368379383414008,
"eval_overall_accuracy": 0.4522023518880208,
"eval_runtime": 1.7594,
"eval_samples_per_second": 8.526,
"eval_steps_per_second": 0.568,
"step": 10
},
{
"epoch": 3.67,
"learning_rate": 5.500000000000001e-06,
"loss": 1.1056,
"step": 11
},
{
"epoch": 4.0,
"learning_rate": 6e-06,
"loss": 1.1072,
"step": 12
},
{
"epoch": 4.33,
"learning_rate": 6.5000000000000004e-06,
"loss": 1.1039,
"step": 13
},
{
"epoch": 4.67,
"learning_rate": 7e-06,
"loss": 1.1023,
"step": 14
},
{
"epoch": 5.0,
"learning_rate": 7.5e-06,
"loss": 1.0945,
"step": 15
},
{
"epoch": 5.33,
"learning_rate": 8e-06,
"loss": 1.0956,
"step": 16
},
{
"epoch": 5.67,
"learning_rate": 8.5e-06,
"loss": 1.0946,
"step": 17
},
{
"epoch": 6.0,
"learning_rate": 9e-06,
"loss": 1.1014,
"step": 18
},
{
"epoch": 6.33,
"learning_rate": 8.973684210526317e-06,
"loss": 1.0905,
"step": 19
},
{
"epoch": 6.67,
"learning_rate": 8.947368421052632e-06,
"loss": 1.0856,
"step": 20
},
{
"epoch": 6.67,
"eval_accuracy_dropoff": 0.5153065913111402,
"eval_accuracy_undropoff": 0.5019142896372908,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.07605219279767873,
"eval_iou_undropoff": 0.49721567971024616,
"eval_iou_unlabeled": 0.0,
"eval_loss": 1.076412320137024,
"eval_mean_accuracy": 0.5086104404742156,
"eval_mean_iou": 0.19108929083597495,
"eval_overall_accuracy": 0.5025072733561198,
"eval_runtime": 1.9083,
"eval_samples_per_second": 7.861,
"eval_steps_per_second": 0.524,
"step": 20
},
{
"epoch": 7.0,
"learning_rate": 8.921052631578947e-06,
"loss": 1.0941,
"step": 21
},
{
"epoch": 7.33,
"learning_rate": 8.894736842105264e-06,
"loss": 1.0807,
"step": 22
},
{
"epoch": 7.67,
"learning_rate": 8.86842105263158e-06,
"loss": 1.0792,
"step": 23
},
{
"epoch": 8.0,
"learning_rate": 8.842105263157895e-06,
"loss": 1.0701,
"step": 24
},
{
"epoch": 8.33,
"learning_rate": 8.81578947368421e-06,
"loss": 1.0735,
"step": 25
},
{
"epoch": 8.67,
"learning_rate": 8.789473684210527e-06,
"loss": 1.0715,
"step": 26
},
{
"epoch": 9.0,
"learning_rate": 8.763157894736842e-06,
"loss": 1.0614,
"step": 27
},
{
"epoch": 9.33,
"learning_rate": 8.736842105263158e-06,
"loss": 1.061,
"step": 28
},
{
"epoch": 9.67,
"learning_rate": 8.710526315789473e-06,
"loss": 1.0587,
"step": 29
},
{
"epoch": 10.0,
"learning_rate": 8.68421052631579e-06,
"loss": 1.0755,
"step": 30
},
{
"epoch": 10.0,
"eval_accuracy_dropoff": 0.704510993176649,
"eval_accuracy_undropoff": 0.5688494996876041,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.11037454693354545,
"eval_iou_undropoff": 0.565151253990595,
"eval_iou_unlabeled": 0.0,
"eval_loss": 1.0611040592193604,
"eval_mean_accuracy": 0.6366802464321266,
"eval_mean_iou": 0.22517526697471346,
"eval_overall_accuracy": 0.5748563130696615,
"eval_runtime": 1.6608,
"eval_samples_per_second": 9.032,
"eval_steps_per_second": 0.602,
"step": 30
},
{
"epoch": 10.33,
"learning_rate": 8.657894736842105e-06,
"loss": 1.0575,
"step": 31
},
{
"epoch": 10.67,
"learning_rate": 8.63157894736842e-06,
"loss": 1.0559,
"step": 32
},
{
"epoch": 11.0,
"learning_rate": 8.605263157894737e-06,
"loss": 1.0358,
"step": 33
},
{
"epoch": 11.33,
"learning_rate": 8.578947368421053e-06,
"loss": 1.0525,
"step": 34
},
{
"epoch": 11.67,
"learning_rate": 8.552631578947368e-06,
"loss": 1.0447,
"step": 35
},
{
"epoch": 12.0,
"learning_rate": 8.526315789473683e-06,
"loss": 1.0542,
"step": 36
},
{
"epoch": 12.33,
"learning_rate": 8.5e-06,
"loss": 1.04,
"step": 37
},
{
"epoch": 12.67,
"learning_rate": 8.473684210526317e-06,
"loss": 1.0351,
"step": 38
},
{
"epoch": 13.0,
"learning_rate": 8.447368421052632e-06,
"loss": 1.0672,
"step": 39
},
{
"epoch": 13.33,
"learning_rate": 8.421052631578947e-06,
"loss": 1.0285,
"step": 40
},
{
"epoch": 13.33,
"eval_accuracy_dropoff": 0.8494440232499368,
"eval_accuracy_undropoff": 0.6478915140077891,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.14199210424243355,
"eval_iou_undropoff": 0.6445086621047246,
"eval_iou_unlabeled": 0.0,
"eval_loss": 1.038216471672058,
"eval_mean_accuracy": 0.748667768628863,
"eval_mean_iou": 0.2621669221157194,
"eval_overall_accuracy": 0.6568158467610677,
"eval_runtime": 1.6441,
"eval_samples_per_second": 9.123,
"eval_steps_per_second": 0.608,
"step": 40
},
{
"epoch": 13.67,
"learning_rate": 8.394736842105263e-06,
"loss": 1.0311,
"step": 41
},
{
"epoch": 14.0,
"learning_rate": 8.36842105263158e-06,
"loss": 1.0374,
"step": 42
},
{
"epoch": 14.33,
"learning_rate": 8.342105263157895e-06,
"loss": 1.0281,
"step": 43
},
{
"epoch": 14.67,
"learning_rate": 8.31578947368421e-06,
"loss": 1.0238,
"step": 44
},
{
"epoch": 15.0,
"learning_rate": 8.289473684210526e-06,
"loss": 0.9876,
"step": 45
},
{
"epoch": 15.33,
"learning_rate": 8.263157894736843e-06,
"loss": 1.0185,
"step": 46
},
{
"epoch": 15.67,
"learning_rate": 8.236842105263158e-06,
"loss": 1.0069,
"step": 47
},
{
"epoch": 16.0,
"learning_rate": 8.210526315789473e-06,
"loss": 1.0178,
"step": 48
},
{
"epoch": 16.33,
"learning_rate": 8.18421052631579e-06,
"loss": 1.014,
"step": 49
},
{
"epoch": 16.67,
"learning_rate": 8.157894736842105e-06,
"loss": 0.9935,
"step": 50
},
{
"epoch": 16.67,
"eval_accuracy_dropoff": 0.8486341810830059,
"eval_accuracy_undropoff": 0.7141391337852696,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.15795859939683365,
"eval_iou_undropoff": 0.7099499580316209,
"eval_iou_unlabeled": 0.0,
"eval_loss": 1.0151127576828003,
"eval_mean_accuracy": 0.7813866574341377,
"eval_mean_iou": 0.2893028524761515,
"eval_overall_accuracy": 0.7200942993164062,
"eval_runtime": 1.8541,
"eval_samples_per_second": 8.09,
"eval_steps_per_second": 0.539,
"step": 50
},
{
"epoch": 17.0,
"learning_rate": 8.131578947368421e-06,
"loss": 0.9613,
"step": 51
},
{
"epoch": 17.33,
"learning_rate": 8.105263157894736e-06,
"loss": 0.9968,
"step": 52
},
{
"epoch": 17.67,
"learning_rate": 8.078947368421053e-06,
"loss": 0.9971,
"step": 53
},
{
"epoch": 18.0,
"learning_rate": 8.052631578947368e-06,
"loss": 0.9627,
"step": 54
},
{
"epoch": 18.33,
"learning_rate": 8.026315789473685e-06,
"loss": 0.9881,
"step": 55
},
{
"epoch": 18.67,
"learning_rate": 8e-06,
"loss": 0.9927,
"step": 56
},
{
"epoch": 19.0,
"learning_rate": 7.973684210526316e-06,
"loss": 1.0041,
"step": 57
},
{
"epoch": 19.33,
"learning_rate": 7.947368421052631e-06,
"loss": 0.9613,
"step": 58
},
{
"epoch": 19.67,
"learning_rate": 7.921052631578948e-06,
"loss": 0.987,
"step": 59
},
{
"epoch": 20.0,
"learning_rate": 7.894736842105263e-06,
"loss": 0.9927,
"step": 60
},
{
"epoch": 20.0,
"eval_accuracy_dropoff": 0.8124325131527558,
"eval_accuracy_undropoff": 0.7801403493086312,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.17348479366555017,
"eval_iou_undropoff": 0.774391632219126,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.9833822846412659,
"eval_mean_accuracy": 0.7962864312306934,
"eval_mean_iou": 0.3159588086282254,
"eval_overall_accuracy": 0.7815701802571614,
"eval_runtime": 1.8711,
"eval_samples_per_second": 8.017,
"eval_steps_per_second": 0.534,
"step": 60
},
{
"epoch": 20.33,
"learning_rate": 7.86842105263158e-06,
"loss": 0.9642,
"step": 61
},
{
"epoch": 20.67,
"learning_rate": 7.842105263157895e-06,
"loss": 0.9628,
"step": 62
},
{
"epoch": 21.0,
"learning_rate": 7.815789473684211e-06,
"loss": 1.0234,
"step": 63
},
{
"epoch": 21.33,
"learning_rate": 7.789473684210526e-06,
"loss": 0.9644,
"step": 64
},
{
"epoch": 21.67,
"learning_rate": 7.763157894736843e-06,
"loss": 0.9831,
"step": 65
},
{
"epoch": 22.0,
"learning_rate": 7.736842105263158e-06,
"loss": 0.9766,
"step": 66
},
{
"epoch": 22.33,
"learning_rate": 7.710526315789474e-06,
"loss": 0.9637,
"step": 67
},
{
"epoch": 22.67,
"learning_rate": 7.68421052631579e-06,
"loss": 0.9429,
"step": 68
},
{
"epoch": 23.0,
"learning_rate": 7.657894736842106e-06,
"loss": 1.0102,
"step": 69
},
{
"epoch": 23.33,
"learning_rate": 7.631578947368421e-06,
"loss": 0.938,
"step": 70
},
{
"epoch": 23.33,
"eval_accuracy_dropoff": 0.8036276334229329,
"eval_accuracy_undropoff": 0.8131103561100272,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.18604038910547227,
"eval_iou_undropoff": 0.8064630295685331,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.9585382342338562,
"eval_mean_accuracy": 0.80836899476648,
"eval_mean_iou": 0.3308344728913351,
"eval_overall_accuracy": 0.8126904805501302,
"eval_runtime": 1.7808,
"eval_samples_per_second": 8.423,
"eval_steps_per_second": 0.562,
"step": 70
},
{
"epoch": 23.67,
"learning_rate": 7.605263157894738e-06,
"loss": 0.9383,
"step": 71
},
{
"epoch": 24.0,
"learning_rate": 7.578947368421053e-06,
"loss": 0.9818,
"step": 72
},
{
"epoch": 24.33,
"learning_rate": 7.5526315789473686e-06,
"loss": 0.9393,
"step": 73
},
{
"epoch": 24.67,
"learning_rate": 7.526315789473684e-06,
"loss": 0.9402,
"step": 74
},
{
"epoch": 25.0,
"learning_rate": 7.5e-06,
"loss": 1.0092,
"step": 75
},
{
"epoch": 25.33,
"learning_rate": 7.473684210526315e-06,
"loss": 0.9253,
"step": 76
},
{
"epoch": 25.67,
"learning_rate": 7.447368421052632e-06,
"loss": 0.928,
"step": 77
},
{
"epoch": 26.0,
"learning_rate": 7.421052631578948e-06,
"loss": 1.0012,
"step": 78
},
{
"epoch": 26.33,
"learning_rate": 7.3947368421052635e-06,
"loss": 0.9346,
"step": 79
},
{
"epoch": 26.67,
"learning_rate": 7.3684210526315784e-06,
"loss": 0.9169,
"step": 80
},
{
"epoch": 26.67,
"eval_accuracy_dropoff": 0.7942541411078181,
"eval_accuracy_undropoff": 0.8396427191534337,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.2048256657878366,
"eval_iou_undropoff": 0.8323542934650573,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.9375818371772766,
"eval_mean_accuracy": 0.8169484301306259,
"eval_mean_iou": 0.34572665308429795,
"eval_overall_accuracy": 0.8376330057779948,
"eval_runtime": 1.7816,
"eval_samples_per_second": 8.419,
"eval_steps_per_second": 0.561,
"step": 80
},
{
"epoch": 27.0,
"learning_rate": 7.342105263157895e-06,
"loss": 0.9599,
"step": 81
},
{
"epoch": 27.33,
"learning_rate": 7.315789473684211e-06,
"loss": 0.9316,
"step": 82
},
{
"epoch": 27.67,
"learning_rate": 7.289473684210527e-06,
"loss": 0.9335,
"step": 83
},
{
"epoch": 28.0,
"learning_rate": 7.263157894736842e-06,
"loss": 0.8321,
"step": 84
},
{
"epoch": 28.33,
"learning_rate": 7.2368421052631575e-06,
"loss": 0.9137,
"step": 85
},
{
"epoch": 28.67,
"learning_rate": 7.210526315789474e-06,
"loss": 0.8788,
"step": 86
},
{
"epoch": 29.0,
"learning_rate": 7.18421052631579e-06,
"loss": 0.9947,
"step": 87
},
{
"epoch": 29.33,
"learning_rate": 7.157894736842105e-06,
"loss": 0.903,
"step": 88
},
{
"epoch": 29.67,
"learning_rate": 7.131578947368421e-06,
"loss": 0.8931,
"step": 89
},
{
"epoch": 30.0,
"learning_rate": 7.1052631578947375e-06,
"loss": 0.8814,
"step": 90
},
{
"epoch": 30.0,
"eval_accuracy_dropoff": 0.7421428079123302,
"eval_accuracy_undropoff": 0.8750230172440403,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.2220341575149798,
"eval_iou_undropoff": 0.8651189466137869,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.9003341197967529,
"eval_mean_accuracy": 0.8085829125781853,
"eval_mean_iou": 0.36238436804292223,
"eval_overall_accuracy": 0.8691393534342448,
"eval_runtime": 1.8384,
"eval_samples_per_second": 8.159,
"eval_steps_per_second": 0.544,
"step": 90
},
{
"epoch": 30.33,
"learning_rate": 7.078947368421053e-06,
"loss": 0.8804,
"step": 91
},
{
"epoch": 30.67,
"learning_rate": 7.052631578947368e-06,
"loss": 0.9237,
"step": 92
},
{
"epoch": 31.0,
"learning_rate": 7.026315789473684e-06,
"loss": 0.9738,
"step": 93
},
{
"epoch": 31.33,
"learning_rate": 7e-06,
"loss": 0.8767,
"step": 94
},
{
"epoch": 31.67,
"learning_rate": 6.9736842105263166e-06,
"loss": 0.8922,
"step": 95
},
{
"epoch": 32.0,
"learning_rate": 6.9473684210526315e-06,
"loss": 0.895,
"step": 96
},
{
"epoch": 32.33,
"learning_rate": 6.921052631578947e-06,
"loss": 0.8999,
"step": 97
},
{
"epoch": 32.67,
"learning_rate": 6.894736842105263e-06,
"loss": 0.8614,
"step": 98
},
{
"epoch": 33.0,
"learning_rate": 6.86842105263158e-06,
"loss": 0.9943,
"step": 99
},
{
"epoch": 33.33,
"learning_rate": 6.842105263157895e-06,
"loss": 0.8618,
"step": 100
},
{
"epoch": 33.33,
"eval_accuracy_dropoff": 0.7550371034070806,
"eval_accuracy_undropoff": 0.8817376662164335,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.2286707289707921,
"eval_iou_undropoff": 0.8719697906661579,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.8893593549728394,
"eval_mean_accuracy": 0.8183873848117571,
"eval_mean_iou": 0.36688017321231664,
"eval_overall_accuracy": 0.8761276245117188,
"eval_runtime": 1.9949,
"eval_samples_per_second": 7.519,
"eval_steps_per_second": 0.501,
"step": 100
},
{
"epoch": 33.67,
"learning_rate": 6.815789473684211e-06,
"loss": 0.875,
"step": 101
},
{
"epoch": 34.0,
"learning_rate": 6.7894736842105264e-06,
"loss": 0.9501,
"step": 102
},
{
"epoch": 34.33,
"learning_rate": 6.763157894736842e-06,
"loss": 0.8395,
"step": 103
},
{
"epoch": 34.67,
"learning_rate": 6.736842105263158e-06,
"loss": 0.8817,
"step": 104
},
{
"epoch": 35.0,
"learning_rate": 6.710526315789474e-06,
"loss": 0.8015,
"step": 105
},
{
"epoch": 35.33,
"learning_rate": 6.68421052631579e-06,
"loss": 0.8708,
"step": 106
},
{
"epoch": 35.67,
"learning_rate": 6.6578947368421055e-06,
"loss": 0.823,
"step": 107
},
{
"epoch": 36.0,
"learning_rate": 6.6315789473684205e-06,
"loss": 0.9683,
"step": 108
},
{
"epoch": 36.33,
"learning_rate": 6.605263157894737e-06,
"loss": 0.86,
"step": 109
},
{
"epoch": 36.67,
"learning_rate": 6.578947368421053e-06,
"loss": 0.8388,
"step": 110
},
{
"epoch": 36.67,
"eval_accuracy_dropoff": 0.718686102878673,
"eval_accuracy_undropoff": 0.9006134029012903,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.24306521418192836,
"eval_iou_undropoff": 0.8892370991006613,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.8618280291557312,
"eval_mean_accuracy": 0.8096497528899816,
"eval_mean_iou": 0.37743410442752984,
"eval_overall_accuracy": 0.8925580342610677,
"eval_runtime": 1.913,
"eval_samples_per_second": 7.841,
"eval_steps_per_second": 0.523,
"step": 110
},
{
"epoch": 37.0,
"learning_rate": 6.552631578947369e-06,
"loss": 0.8596,
"step": 111
},
{
"epoch": 37.33,
"learning_rate": 6.526315789473684e-06,
"loss": 0.8692,
"step": 112
},
{
"epoch": 37.67,
"learning_rate": 6.5000000000000004e-06,
"loss": 0.8117,
"step": 113
},
{
"epoch": 38.0,
"learning_rate": 6.473684210526316e-06,
"loss": 0.9507,
"step": 114
},
{
"epoch": 38.33,
"learning_rate": 6.447368421052632e-06,
"loss": 0.8147,
"step": 115
},
{
"epoch": 38.67,
"learning_rate": 6.421052631578947e-06,
"loss": 0.8928,
"step": 116
},
{
"epoch": 39.0,
"learning_rate": 6.394736842105263e-06,
"loss": 0.826,
"step": 117
},
{
"epoch": 39.33,
"learning_rate": 6.3684210526315795e-06,
"loss": 0.8085,
"step": 118
},
{
"epoch": 39.67,
"learning_rate": 6.342105263157895e-06,
"loss": 0.8143,
"step": 119
},
{
"epoch": 40.0,
"learning_rate": 6.31578947368421e-06,
"loss": 0.8878,
"step": 120
},
{
"epoch": 40.0,
"eval_accuracy_dropoff": 0.6618248443494842,
"eval_accuracy_undropoff": 0.9256646794669153,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.2670577274286403,
"eval_iou_undropoff": 0.9115604296316833,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.8268659710884094,
"eval_mean_accuracy": 0.7937447619081998,
"eval_mean_iou": 0.39287271902010784,
"eval_overall_accuracy": 0.9139823913574219,
"eval_runtime": 1.8743,
"eval_samples_per_second": 8.003,
"eval_steps_per_second": 0.534,
"step": 120
},
{
"epoch": 40.33,
"learning_rate": 6.289473684210526e-06,
"loss": 0.8174,
"step": 121
},
{
"epoch": 40.67,
"learning_rate": 6.263157894736843e-06,
"loss": 0.8032,
"step": 122
},
{
"epoch": 41.0,
"learning_rate": 6.236842105263159e-06,
"loss": 0.9556,
"step": 123
},
{
"epoch": 41.33,
"learning_rate": 6.210526315789474e-06,
"loss": 0.795,
"step": 124
},
{
"epoch": 41.67,
"learning_rate": 6.184210526315789e-06,
"loss": 0.8119,
"step": 125
},
{
"epoch": 42.0,
"learning_rate": 6.157894736842105e-06,
"loss": 0.8237,
"step": 126
},
{
"epoch": 42.33,
"learning_rate": 6.131578947368422e-06,
"loss": 0.7689,
"step": 127
},
{
"epoch": 42.67,
"learning_rate": 6.105263157894737e-06,
"loss": 0.8232,
"step": 128
},
{
"epoch": 43.0,
"learning_rate": 6.078947368421053e-06,
"loss": 0.9406,
"step": 129
},
{
"epoch": 43.33,
"learning_rate": 6.0526315789473685e-06,
"loss": 0.8066,
"step": 130
},
{
"epoch": 43.33,
"eval_accuracy_dropoff": 0.6561502056195005,
"eval_accuracy_undropoff": 0.9348247443090196,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.28393932540146094,
"eval_iou_undropoff": 0.9202498341878032,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.8073786497116089,
"eval_mean_accuracy": 0.7954874749642601,
"eval_mean_iou": 0.4013963865297547,
"eval_overall_accuracy": 0.9224856058756511,
"eval_runtime": 1.8279,
"eval_samples_per_second": 8.206,
"eval_steps_per_second": 0.547,
"step": 130
},
{
"epoch": 43.67,
"learning_rate": 6.026315789473685e-06,
"loss": 0.756,
"step": 131
},
{
"epoch": 44.0,
"learning_rate": 6e-06,
"loss": 0.9245,
"step": 132
},
{
"epoch": 44.33,
"learning_rate": 5.973684210526316e-06,
"loss": 0.7635,
"step": 133
},
{
"epoch": 44.67,
"learning_rate": 5.947368421052632e-06,
"loss": 0.8357,
"step": 134
},
{
"epoch": 45.0,
"learning_rate": 5.921052631578948e-06,
"loss": 0.8333,
"step": 135
},
{
"epoch": 45.33,
"learning_rate": 5.894736842105263e-06,
"loss": 0.799,
"step": 136
},
{
"epoch": 45.67,
"learning_rate": 5.868421052631579e-06,
"loss": 0.7785,
"step": 137
},
{
"epoch": 46.0,
"learning_rate": 5.842105263157895e-06,
"loss": 0.7472,
"step": 138
},
{
"epoch": 46.33,
"learning_rate": 5.815789473684211e-06,
"loss": 0.7746,
"step": 139
},
{
"epoch": 46.67,
"learning_rate": 5.789473684210526e-06,
"loss": 0.8084,
"step": 140
},
{
"epoch": 46.67,
"eval_accuracy_dropoff": 0.6487237806419004,
"eval_accuracy_undropoff": 0.9376131570292269,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.2843927433873425,
"eval_iou_undropoff": 0.922643123497163,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.7918700575828552,
"eval_mean_accuracy": 0.7931684688355636,
"eval_mean_iou": 0.40234528896150185,
"eval_overall_accuracy": 0.9248217264811198,
"eval_runtime": 1.8666,
"eval_samples_per_second": 8.036,
"eval_steps_per_second": 0.536,
"step": 140
},
{
"epoch": 47.0,
"learning_rate": 5.7631578947368425e-06,
"loss": 0.9072,
"step": 141
},
{
"epoch": 47.33,
"learning_rate": 5.736842105263158e-06,
"loss": 0.7663,
"step": 142
},
{
"epoch": 47.67,
"learning_rate": 5.710526315789474e-06,
"loss": 0.7524,
"step": 143
},
{
"epoch": 48.0,
"learning_rate": 5.684210526315789e-06,
"loss": 0.9309,
"step": 144
},
{
"epoch": 48.33,
"learning_rate": 5.657894736842105e-06,
"loss": 0.7694,
"step": 145
},
{
"epoch": 48.67,
"learning_rate": 5.631578947368422e-06,
"loss": 0.7437,
"step": 146
},
{
"epoch": 49.0,
"learning_rate": 5.6052631578947374e-06,
"loss": 0.8711,
"step": 147
},
{
"epoch": 49.33,
"learning_rate": 5.578947368421052e-06,
"loss": 0.7414,
"step": 148
},
{
"epoch": 49.67,
"learning_rate": 5.552631578947368e-06,
"loss": 0.7699,
"step": 149
},
{
"epoch": 50.0,
"learning_rate": 5.526315789473685e-06,
"loss": 0.7415,
"step": 150
},
{
"epoch": 50.0,
"eval_accuracy_dropoff": 0.6249109747972523,
"eval_accuracy_undropoff": 0.9451210361112619,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.29126956931445824,
"eval_iou_undropoff": 0.9289919517149979,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.7706870436668396,
"eval_mean_accuracy": 0.7850160054542571,
"eval_mean_iou": 0.4067538403431521,
"eval_overall_accuracy": 0.9309427897135417,
"eval_runtime": 1.868,
"eval_samples_per_second": 8.03,
"eval_steps_per_second": 0.535,
"step": 150
},
{
"epoch": 50.33,
"learning_rate": 5.500000000000001e-06,
"loss": 0.7662,
"step": 151
},
{
"epoch": 50.67,
"learning_rate": 5.473684210526316e-06,
"loss": 0.735,
"step": 152
},
{
"epoch": 51.0,
"learning_rate": 5.4473684210526315e-06,
"loss": 0.7445,
"step": 153
},
{
"epoch": 51.33,
"learning_rate": 5.421052631578947e-06,
"loss": 0.7691,
"step": 154
},
{
"epoch": 51.67,
"learning_rate": 5.394736842105264e-06,
"loss": 0.7024,
"step": 155
},
{
"epoch": 52.0,
"learning_rate": 5.368421052631579e-06,
"loss": 0.9282,
"step": 156
},
{
"epoch": 52.33,
"learning_rate": 5.342105263157895e-06,
"loss": 0.7189,
"step": 157
},
{
"epoch": 52.67,
"learning_rate": 5.315789473684211e-06,
"loss": 0.7815,
"step": 158
},
{
"epoch": 53.0,
"learning_rate": 5.289473684210527e-06,
"loss": 0.7113,
"step": 159
},
{
"epoch": 53.33,
"learning_rate": 5.263157894736842e-06,
"loss": 0.7508,
"step": 160
},
{
"epoch": 53.33,
"eval_accuracy_dropoff": 0.5735348174696165,
"eval_accuracy_undropoff": 0.9585285142408886,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.30633518727011133,
"eval_iou_undropoff": 0.9399596899192442,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.7325581312179565,
"eval_mean_accuracy": 0.7660316658552526,
"eval_mean_iou": 0.4154316257297852,
"eval_overall_accuracy": 0.9414817810058593,
"eval_runtime": 1.8819,
"eval_samples_per_second": 7.971,
"eval_steps_per_second": 0.531,
"step": 160
},
{
"epoch": 53.67,
"learning_rate": 5.236842105263158e-06,
"loss": 0.7154,
"step": 161
},
{
"epoch": 54.0,
"learning_rate": 5.210526315789474e-06,
"loss": 0.9285,
"step": 162
},
{
"epoch": 54.33,
"learning_rate": 5.18421052631579e-06,
"loss": 0.7292,
"step": 163
},
{
"epoch": 54.67,
"learning_rate": 5.1578947368421055e-06,
"loss": 0.7165,
"step": 164
},
{
"epoch": 55.0,
"learning_rate": 5.131578947368421e-06,
"loss": 0.743,
"step": 165
},
{
"epoch": 55.33,
"learning_rate": 5.105263157894737e-06,
"loss": 0.7176,
"step": 166
},
{
"epoch": 55.67,
"learning_rate": 5.078947368421053e-06,
"loss": 0.7,
"step": 167
},
{
"epoch": 56.0,
"learning_rate": 5.052631578947368e-06,
"loss": 0.9033,
"step": 168
},
{
"epoch": 56.33,
"learning_rate": 5.026315789473685e-06,
"loss": 0.6789,
"step": 169
},
{
"epoch": 56.67,
"learning_rate": 5e-06,
"loss": 0.7312,
"step": 170
},
{
"epoch": 56.67,
"eval_accuracy_dropoff": 0.5645863487031039,
"eval_accuracy_undropoff": 0.9625257979399966,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.31545015483850264,
"eval_iou_undropoff": 0.943493256536948,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.7126114368438721,
"eval_mean_accuracy": 0.7635560733215503,
"eval_mean_iou": 0.41964780379181693,
"eval_overall_accuracy": 0.9449058532714844,
"eval_runtime": 1.9872,
"eval_samples_per_second": 7.548,
"eval_steps_per_second": 0.503,
"step": 170
},
{
"epoch": 57.0,
"learning_rate": 4.973684210526316e-06,
"loss": 0.9239,
"step": 171
},
{
"epoch": 57.33,
"learning_rate": 4.947368421052631e-06,
"loss": 0.7121,
"step": 172
},
{
"epoch": 57.67,
"learning_rate": 4.921052631578948e-06,
"loss": 0.6922,
"step": 173
},
{
"epoch": 58.0,
"learning_rate": 4.894736842105264e-06,
"loss": 0.9187,
"step": 174
},
{
"epoch": 58.33,
"learning_rate": 4.8684210526315795e-06,
"loss": 0.6794,
"step": 175
},
{
"epoch": 58.67,
"learning_rate": 4.8421052631578945e-06,
"loss": 0.7535,
"step": 176
},
{
"epoch": 59.0,
"learning_rate": 4.81578947368421e-06,
"loss": 0.6672,
"step": 177
},
{
"epoch": 59.33,
"learning_rate": 4.789473684210527e-06,
"loss": 0.7101,
"step": 178
},
{
"epoch": 59.67,
"learning_rate": 4.763157894736843e-06,
"loss": 0.6992,
"step": 179
},
{
"epoch": 60.0,
"learning_rate": 4.736842105263158e-06,
"loss": 0.6442,
"step": 180
},
{
"epoch": 60.0,
"eval_accuracy_dropoff": 0.5296252900498541,
"eval_accuracy_undropoff": 0.9704251032183695,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.3267820059394292,
"eval_iou_undropoff": 0.94972848590658,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.6868845820426941,
"eval_mean_accuracy": 0.7500251966341118,
"eval_mean_iou": 0.4255034972820031,
"eval_overall_accuracy": 0.9509073893229166,
"eval_runtime": 1.8323,
"eval_samples_per_second": 8.187,
"eval_steps_per_second": 0.546,
"step": 180
},
{
"epoch": 60.33,
"learning_rate": 4.7105263157894736e-06,
"loss": 0.7235,
"step": 181
},
{
"epoch": 60.67,
"learning_rate": 4.68421052631579e-06,
"loss": 0.685,
"step": 182
},
{
"epoch": 61.0,
"learning_rate": 4.657894736842106e-06,
"loss": 0.7571,
"step": 183
},
{
"epoch": 61.33,
"learning_rate": 4.631578947368421e-06,
"loss": 0.6767,
"step": 184
},
{
"epoch": 61.67,
"learning_rate": 4.605263157894737e-06,
"loss": 0.7594,
"step": 185
},
{
"epoch": 62.0,
"learning_rate": 4.578947368421053e-06,
"loss": 0.6212,
"step": 186
},
{
"epoch": 62.33,
"learning_rate": 4.552631578947369e-06,
"loss": 0.6557,
"step": 187
},
{
"epoch": 62.67,
"learning_rate": 4.526315789473684e-06,
"loss": 0.7043,
"step": 188
},
{
"epoch": 63.0,
"learning_rate": 4.5e-06,
"loss": 0.8759,
"step": 189
},
{
"epoch": 63.33,
"learning_rate": 4.473684210526316e-06,
"loss": 0.6633,
"step": 190
},
{
"epoch": 63.33,
"eval_accuracy_dropoff": 0.5328129666643693,
"eval_accuracy_undropoff": 0.9719144386506626,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.33433163945651784,
"eval_iou_undropoff": 0.9513240511320675,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.6765334010124207,
"eval_mean_accuracy": 0.752363702657516,
"eval_mean_iou": 0.4285518968628618,
"eval_overall_accuracy": 0.952471923828125,
"eval_runtime": 1.8425,
"eval_samples_per_second": 8.141,
"eval_steps_per_second": 0.543,
"step": 190
},
{
"epoch": 63.67,
"learning_rate": 4.447368421052632e-06,
"loss": 0.729,
"step": 191
},
{
"epoch": 64.0,
"learning_rate": 4.4210526315789476e-06,
"loss": 0.8651,
"step": 192
},
{
"epoch": 64.33,
"learning_rate": 4.394736842105263e-06,
"loss": 0.6915,
"step": 193
},
{
"epoch": 64.67,
"learning_rate": 4.368421052631579e-06,
"loss": 0.6817,
"step": 194
},
{
"epoch": 65.0,
"learning_rate": 4.342105263157895e-06,
"loss": 0.6375,
"step": 195
},
{
"epoch": 65.33,
"learning_rate": 4.31578947368421e-06,
"loss": 0.6665,
"step": 196
},
{
"epoch": 65.67,
"learning_rate": 4.289473684210527e-06,
"loss": 0.6473,
"step": 197
},
{
"epoch": 66.0,
"learning_rate": 4.263157894736842e-06,
"loss": 0.6426,
"step": 198
},
{
"epoch": 66.33,
"learning_rate": 4.236842105263158e-06,
"loss": 0.6096,
"step": 199
},
{
"epoch": 66.67,
"learning_rate": 4.210526315789473e-06,
"loss": 0.7247,
"step": 200
},
{
"epoch": 66.67,
"eval_accuracy_dropoff": 0.4885990304868243,
"eval_accuracy_undropoff": 0.9784518681487111,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.33639268286896074,
"eval_iou_undropoff": 0.9558068336362464,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.6556523442268372,
"eval_mean_accuracy": 0.7335254493177676,
"eval_mean_iou": 0.4307331721684024,
"eval_overall_accuracy": 0.956762186686198,
"eval_runtime": 1.8078,
"eval_samples_per_second": 8.297,
"eval_steps_per_second": 0.553,
"step": 200
},
{
"epoch": 67.0,
"learning_rate": 4.18421052631579e-06,
"loss": 0.8805,
"step": 201
},
{
"epoch": 67.33,
"learning_rate": 4.157894736842105e-06,
"loss": 0.6557,
"step": 202
},
{
"epoch": 67.67,
"learning_rate": 4.1315789473684216e-06,
"loss": 0.6537,
"step": 203
},
{
"epoch": 68.0,
"learning_rate": 4.1052631578947365e-06,
"loss": 0.5931,
"step": 204
},
{
"epoch": 68.33,
"learning_rate": 4.078947368421052e-06,
"loss": 0.6597,
"step": 205
},
{
"epoch": 68.67,
"learning_rate": 4.052631578947368e-06,
"loss": 0.6736,
"step": 206
},
{
"epoch": 69.0,
"learning_rate": 4.026315789473684e-06,
"loss": 0.8674,
"step": 207
},
{
"epoch": 69.33,
"learning_rate": 4e-06,
"loss": 0.6612,
"step": 208
},
{
"epoch": 69.67,
"learning_rate": 3.973684210526316e-06,
"loss": 0.6979,
"step": 209
},
{
"epoch": 70.0,
"learning_rate": 3.9473684210526315e-06,
"loss": 0.6133,
"step": 210
},
{
"epoch": 70.0,
"eval_accuracy_dropoff": 0.4761240149792083,
"eval_accuracy_undropoff": 0.9795952796821332,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.3330293510312633,
"eval_iou_undropoff": 0.9563835759643214,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.636894166469574,
"eval_mean_accuracy": 0.7278596473306708,
"eval_mean_iou": 0.42980430899852823,
"eval_overall_accuracy": 0.9573026021321615,
"eval_runtime": 1.8877,
"eval_samples_per_second": 7.946,
"eval_steps_per_second": 0.53,
"step": 210
},
{
"epoch": 70.33,
"learning_rate": 3.921052631578947e-06,
"loss": 0.6461,
"step": 211
},
{
"epoch": 70.67,
"learning_rate": 3.894736842105263e-06,
"loss": 0.6805,
"step": 212
},
{
"epoch": 71.0,
"learning_rate": 3.868421052631579e-06,
"loss": 0.8854,
"step": 213
},
{
"epoch": 71.33,
"learning_rate": 3.842105263157895e-06,
"loss": 0.6374,
"step": 214
},
{
"epoch": 71.67,
"learning_rate": 3.8157894736842105e-06,
"loss": 0.6382,
"step": 215
},
{
"epoch": 72.0,
"learning_rate": 3.7894736842105264e-06,
"loss": 0.8767,
"step": 216
},
{
"epoch": 72.33,
"learning_rate": 3.763157894736842e-06,
"loss": 0.641,
"step": 217
},
{
"epoch": 72.67,
"learning_rate": 3.7368421052631576e-06,
"loss": 0.6602,
"step": 218
},
{
"epoch": 73.0,
"learning_rate": 3.710526315789474e-06,
"loss": 0.8691,
"step": 219
},
{
"epoch": 73.33,
"learning_rate": 3.6842105263157892e-06,
"loss": 0.6309,
"step": 220
},
{
"epoch": 73.33,
"eval_accuracy_dropoff": 0.5122797344177178,
"eval_accuracy_undropoff": 0.9751879431152097,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.33562496942603737,
"eval_iou_undropoff": 0.9536397388684995,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.6308692097663879,
"eval_mean_accuracy": 0.7437338387664638,
"eval_mean_iou": 0.4297549027648457,
"eval_overall_accuracy": 0.9546913146972656,
"eval_runtime": 1.8683,
"eval_samples_per_second": 8.029,
"eval_steps_per_second": 0.535,
"step": 220
},
{
"epoch": 73.67,
"learning_rate": 3.6578947368421055e-06,
"loss": 0.6259,
"step": 221
},
{
"epoch": 74.0,
"learning_rate": 3.631578947368421e-06,
"loss": 0.6131,
"step": 222
},
{
"epoch": 74.33,
"learning_rate": 3.605263157894737e-06,
"loss": 0.6548,
"step": 223
},
{
"epoch": 74.67,
"learning_rate": 3.5789473684210525e-06,
"loss": 0.6167,
"step": 224
},
{
"epoch": 75.0,
"learning_rate": 3.5526315789473687e-06,
"loss": 0.5981,
"step": 225
},
{
"epoch": 75.33,
"learning_rate": 3.526315789473684e-06,
"loss": 0.6625,
"step": 226
},
{
"epoch": 75.67,
"learning_rate": 3.5e-06,
"loss": 0.6269,
"step": 227
},
{
"epoch": 76.0,
"learning_rate": 3.4736842105263158e-06,
"loss": 0.6141,
"step": 228
},
{
"epoch": 76.33,
"learning_rate": 3.4473684210526316e-06,
"loss": 0.6309,
"step": 229
},
{
"epoch": 76.67,
"learning_rate": 3.4210526315789474e-06,
"loss": 0.6373,
"step": 230
},
{
"epoch": 76.67,
"eval_accuracy_dropoff": 0.45853148620396533,
"eval_accuracy_undropoff": 0.9807951034206019,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.3261565865376759,
"eval_iou_undropoff": 0.9567931166780796,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.6093744039535522,
"eval_mean_accuracy": 0.7196632948122836,
"eval_mean_iou": 0.4276499010719185,
"eval_overall_accuracy": 0.9576703389485677,
"eval_runtime": 1.8002,
"eval_samples_per_second": 8.332,
"eval_steps_per_second": 0.555,
"step": 230
},
{
"epoch": 77.0,
"learning_rate": 3.3947368421052632e-06,
"loss": 0.6731,
"step": 231
},
{
"epoch": 77.33,
"learning_rate": 3.368421052631579e-06,
"loss": 0.6215,
"step": 232
},
{
"epoch": 77.67,
"learning_rate": 3.342105263157895e-06,
"loss": 0.6207,
"step": 233
},
{
"epoch": 78.0,
"learning_rate": 3.3157894736842103e-06,
"loss": 0.7978,
"step": 234
},
{
"epoch": 78.33,
"learning_rate": 3.2894736842105265e-06,
"loss": 0.5932,
"step": 235
},
{
"epoch": 78.67,
"learning_rate": 3.263157894736842e-06,
"loss": 0.6747,
"step": 236
},
{
"epoch": 79.0,
"learning_rate": 3.236842105263158e-06,
"loss": 0.5737,
"step": 237
},
{
"epoch": 79.33,
"learning_rate": 3.2105263157894735e-06,
"loss": 0.6148,
"step": 238
},
{
"epoch": 79.67,
"learning_rate": 3.1842105263157898e-06,
"loss": 0.632,
"step": 239
},
{
"epoch": 80.0,
"learning_rate": 3.157894736842105e-06,
"loss": 0.8436,
"step": 240
},
{
"epoch": 80.0,
"eval_accuracy_dropoff": 0.5100971810600317,
"eval_accuracy_undropoff": 0.9775950412607383,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.3462631634351839,
"eval_iou_undropoff": 0.9558991275315168,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.6194807887077332,
"eval_mean_accuracy": 0.7438461111603849,
"eval_mean_iou": 0.43405409698890024,
"eval_overall_accuracy": 0.9568951924641927,
"eval_runtime": 1.8614,
"eval_samples_per_second": 8.058,
"eval_steps_per_second": 0.537,
"step": 240
}
],
"max_steps": 360,
"num_train_epochs": 120,
"total_flos": 4.76793375031296e+16,
"trial_name": null,
"trial_params": null
}