sam1120's picture
Training in progress, step 120
8e316ed
raw
history blame
21.8 kB
{
"best_metric": 0.3625328540802002,
"best_model_checkpoint": "/robodata/smodak/Projects/nspl/scripts/terrainseg/training/models/safety-utcustom-train-SF30-RGBD-b0/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": 5e-06,
"loss": 1.0565,
"step": 1
},
{
"epoch": 1.0,
"learning_rate": 1e-05,
"loss": 1.0485,
"step": 2
},
{
"epoch": 1.5,
"learning_rate": 1.5e-05,
"loss": 1.0559,
"step": 3
},
{
"epoch": 2.0,
"learning_rate": 2e-05,
"loss": 1.0501,
"step": 4
},
{
"epoch": 2.5,
"learning_rate": 2.5e-05,
"loss": 1.0391,
"step": 5
},
{
"epoch": 3.0,
"learning_rate": 3e-05,
"loss": 1.0435,
"step": 6
},
{
"epoch": 3.5,
"learning_rate": 3.5e-05,
"loss": 1.0235,
"step": 7
},
{
"epoch": 4.0,
"learning_rate": 4e-05,
"loss": 1.0155,
"step": 8
},
{
"epoch": 4.5,
"learning_rate": 4.5e-05,
"loss": 1.0107,
"step": 9
},
{
"epoch": 5.0,
"learning_rate": 5e-05,
"loss": 0.9925,
"step": 10
},
{
"epoch": 5.0,
"eval_accuracy_safe": 0.16253187340632969,
"eval_accuracy_unlabeled": NaN,
"eval_accuracy_unsafe": 0.9084522788185269,
"eval_iou_safe": 0.04617705465733694,
"eval_iou_unlabeled": 0.0,
"eval_iou_unsafe": 0.8841402740524446,
"eval_loss": 1.0611684322357178,
"eval_mean_accuracy": 0.5354920761124282,
"eval_mean_iou": 0.3101057762365938,
"eval_overall_accuracy": 0.8847389221191406,
"eval_runtime": 2.0635,
"eval_samples_per_second": 3.877,
"eval_steps_per_second": 0.485,
"step": 10
},
{
"epoch": 5.5,
"learning_rate": 4.973684210526316e-05,
"loss": 0.9711,
"step": 11
},
{
"epoch": 6.0,
"learning_rate": 4.9473684210526315e-05,
"loss": 0.9794,
"step": 12
},
{
"epoch": 6.5,
"learning_rate": 4.921052631578947e-05,
"loss": 0.9568,
"step": 13
},
{
"epoch": 7.0,
"learning_rate": 4.8947368421052635e-05,
"loss": 0.9109,
"step": 14
},
{
"epoch": 7.5,
"learning_rate": 4.868421052631579e-05,
"loss": 0.9157,
"step": 15
},
{
"epoch": 8.0,
"learning_rate": 4.842105263157895e-05,
"loss": 0.8739,
"step": 16
},
{
"epoch": 8.5,
"learning_rate": 4.8157894736842105e-05,
"loss": 0.8838,
"step": 17
},
{
"epoch": 9.0,
"learning_rate": 4.789473684210526e-05,
"loss": 0.9098,
"step": 18
},
{
"epoch": 9.5,
"learning_rate": 4.7631578947368424e-05,
"loss": 0.8599,
"step": 19
},
{
"epoch": 10.0,
"learning_rate": 4.736842105263158e-05,
"loss": 0.8589,
"step": 20
},
{
"epoch": 10.0,
"eval_accuracy_safe": 0.05294735263236838,
"eval_accuracy_unlabeled": NaN,
"eval_accuracy_unsafe": 0.9833241565303213,
"eval_iou_safe": 0.03730081575721712,
"eval_iou_unlabeled": 0.0,
"eval_iou_unsafe": 0.9536688093648232,
"eval_loss": 0.9441336989402771,
"eval_mean_accuracy": 0.5181357545813449,
"eval_mean_iou": 0.33032320837401347,
"eval_overall_accuracy": 0.9537467956542969,
"eval_runtime": 1.9246,
"eval_samples_per_second": 4.157,
"eval_steps_per_second": 0.52,
"step": 20
},
{
"epoch": 10.5,
"learning_rate": 4.7105263157894744e-05,
"loss": 0.8438,
"step": 21
},
{
"epoch": 11.0,
"learning_rate": 4.68421052631579e-05,
"loss": 0.828,
"step": 22
},
{
"epoch": 11.5,
"learning_rate": 4.657894736842106e-05,
"loss": 0.8353,
"step": 23
},
{
"epoch": 12.0,
"learning_rate": 4.6315789473684214e-05,
"loss": 0.8821,
"step": 24
},
{
"epoch": 12.5,
"learning_rate": 4.605263157894737e-05,
"loss": 0.8043,
"step": 25
},
{
"epoch": 13.0,
"learning_rate": 4.5789473684210527e-05,
"loss": 0.7807,
"step": 26
},
{
"epoch": 13.5,
"learning_rate": 4.552631578947369e-05,
"loss": 0.7855,
"step": 27
},
{
"epoch": 14.0,
"learning_rate": 4.5263157894736846e-05,
"loss": 0.7243,
"step": 28
},
{
"epoch": 14.5,
"learning_rate": 4.5e-05,
"loss": 0.7318,
"step": 29
},
{
"epoch": 15.0,
"learning_rate": 4.473684210526316e-05,
"loss": 0.7016,
"step": 30
},
{
"epoch": 15.0,
"eval_accuracy_safe": 0.017234138293085347,
"eval_accuracy_unlabeled": NaN,
"eval_accuracy_unsafe": 0.9965333354346406,
"eval_iou_safe": 0.016899046946699612,
"eval_iou_unlabeled": 0.0,
"eval_iou_unsafe": 0.965381728938365,
"eval_loss": 0.7763598561286926,
"eval_mean_accuracy": 0.506883736863863,
"eval_mean_iou": 0.32742692529502154,
"eval_overall_accuracy": 0.9654006958007812,
"eval_runtime": 1.8724,
"eval_samples_per_second": 4.273,
"eval_steps_per_second": 0.534,
"step": 30
},
{
"epoch": 15.5,
"learning_rate": 4.4473684210526316e-05,
"loss": 0.6872,
"step": 31
},
{
"epoch": 16.0,
"learning_rate": 4.421052631578947e-05,
"loss": 0.6695,
"step": 32
},
{
"epoch": 16.5,
"learning_rate": 4.394736842105263e-05,
"loss": 0.6728,
"step": 33
},
{
"epoch": 17.0,
"learning_rate": 4.368421052631579e-05,
"loss": 0.757,
"step": 34
},
{
"epoch": 17.5,
"learning_rate": 4.342105263157895e-05,
"loss": 0.6676,
"step": 35
},
{
"epoch": 18.0,
"learning_rate": 4.3157894736842105e-05,
"loss": 0.6491,
"step": 36
},
{
"epoch": 18.5,
"learning_rate": 4.289473684210527e-05,
"loss": 0.6541,
"step": 37
},
{
"epoch": 19.0,
"learning_rate": 4.2631578947368425e-05,
"loss": 0.6635,
"step": 38
},
{
"epoch": 19.5,
"learning_rate": 4.236842105263158e-05,
"loss": 0.6657,
"step": 39
},
{
"epoch": 20.0,
"learning_rate": 4.210526315789474e-05,
"loss": 0.6093,
"step": 40
},
{
"epoch": 20.0,
"eval_accuracy_safe": 0.053772311384430775,
"eval_accuracy_unlabeled": NaN,
"eval_accuracy_unsafe": 0.9900870827714798,
"eval_iou_safe": 0.04153054841176062,
"eval_iou_unlabeled": 0.0,
"eval_iou_unsafe": 0.9602530036058077,
"eval_loss": 0.6213392615318298,
"eval_mean_accuracy": 0.5219296970779553,
"eval_mean_iou": 0.33392785067252273,
"eval_overall_accuracy": 0.9603209495544434,
"eval_runtime": 1.9138,
"eval_samples_per_second": 4.18,
"eval_steps_per_second": 0.523,
"step": 40
},
{
"epoch": 20.5,
"learning_rate": 4.18421052631579e-05,
"loss": 0.6155,
"step": 41
},
{
"epoch": 21.0,
"learning_rate": 4.157894736842106e-05,
"loss": 0.6127,
"step": 42
},
{
"epoch": 21.5,
"learning_rate": 4.1315789473684214e-05,
"loss": 0.5792,
"step": 43
},
{
"epoch": 22.0,
"learning_rate": 4.105263157894737e-05,
"loss": 0.5893,
"step": 44
},
{
"epoch": 22.5,
"learning_rate": 4.078947368421053e-05,
"loss": 0.6444,
"step": 45
},
{
"epoch": 23.0,
"learning_rate": 4.0526315789473684e-05,
"loss": 0.5666,
"step": 46
},
{
"epoch": 23.5,
"learning_rate": 4.026315789473684e-05,
"loss": 0.5816,
"step": 47
},
{
"epoch": 24.0,
"learning_rate": 4e-05,
"loss": 0.5644,
"step": 48
},
{
"epoch": 24.5,
"learning_rate": 3.973684210526316e-05,
"loss": 0.5985,
"step": 49
},
{
"epoch": 25.0,
"learning_rate": 3.9473684210526316e-05,
"loss": 0.5281,
"step": 50
},
{
"epoch": 25.0,
"eval_accuracy_safe": 0.04757762111894405,
"eval_accuracy_unlabeled": NaN,
"eval_accuracy_unsafe": 0.9951110130501034,
"eval_iou_safe": 0.041666666666666664,
"eval_iou_unlabeled": 0.0,
"eval_iou_unsafe": 0.9649351951785595,
"eval_loss": 0.5431288480758667,
"eval_mean_accuracy": 0.5213443170845238,
"eval_mean_iou": 0.33553395394840874,
"eval_overall_accuracy": 0.9649882316589355,
"eval_runtime": 1.915,
"eval_samples_per_second": 4.178,
"eval_steps_per_second": 0.522,
"step": 50
},
{
"epoch": 25.5,
"learning_rate": 3.921052631578947e-05,
"loss": 0.5531,
"step": 51
},
{
"epoch": 26.0,
"learning_rate": 3.894736842105263e-05,
"loss": 0.5125,
"step": 52
},
{
"epoch": 26.5,
"learning_rate": 3.868421052631579e-05,
"loss": 0.5446,
"step": 53
},
{
"epoch": 27.0,
"learning_rate": 3.842105263157895e-05,
"loss": 0.6699,
"step": 54
},
{
"epoch": 27.5,
"learning_rate": 3.815789473684211e-05,
"loss": 0.5265,
"step": 55
},
{
"epoch": 28.0,
"learning_rate": 3.789473684210527e-05,
"loss": 0.5342,
"step": 56
},
{
"epoch": 28.5,
"learning_rate": 3.7631578947368425e-05,
"loss": 0.5428,
"step": 57
},
{
"epoch": 29.0,
"learning_rate": 3.736842105263158e-05,
"loss": 0.5174,
"step": 58
},
{
"epoch": 29.5,
"learning_rate": 3.710526315789474e-05,
"loss": 0.5173,
"step": 59
},
{
"epoch": 30.0,
"learning_rate": 3.6842105263157895e-05,
"loss": 0.5077,
"step": 60
},
{
"epoch": 30.0,
"eval_accuracy_safe": 0.05236238188090595,
"eval_accuracy_unlabeled": NaN,
"eval_accuracy_unsafe": 0.9937724146286449,
"eval_iou_safe": 0.0444283241702301,
"eval_iou_unlabeled": 0.0,
"eval_iou_unsafe": 0.963784012789081,
"eval_loss": 0.5042787194252014,
"eval_mean_accuracy": 0.5230673982547754,
"eval_mean_iou": 0.336070778986437,
"eval_overall_accuracy": 0.9638442993164062,
"eval_runtime": 1.8605,
"eval_samples_per_second": 4.3,
"eval_steps_per_second": 0.537,
"step": 60
},
{
"epoch": 30.5,
"learning_rate": 3.657894736842106e-05,
"loss": 0.5175,
"step": 61
},
{
"epoch": 31.0,
"learning_rate": 3.6315789473684214e-05,
"loss": 0.5155,
"step": 62
},
{
"epoch": 31.5,
"learning_rate": 3.605263157894737e-05,
"loss": 0.5188,
"step": 63
},
{
"epoch": 32.0,
"learning_rate": 3.578947368421053e-05,
"loss": 0.4911,
"step": 64
},
{
"epoch": 32.5,
"learning_rate": 3.5526315789473684e-05,
"loss": 0.4856,
"step": 65
},
{
"epoch": 33.0,
"learning_rate": 3.526315789473684e-05,
"loss": 0.5214,
"step": 66
},
{
"epoch": 33.5,
"learning_rate": 3.5e-05,
"loss": 0.4863,
"step": 67
},
{
"epoch": 34.0,
"learning_rate": 3.473684210526316e-05,
"loss": 0.5344,
"step": 68
},
{
"epoch": 34.5,
"learning_rate": 3.447368421052632e-05,
"loss": 0.4722,
"step": 69
},
{
"epoch": 35.0,
"learning_rate": 3.421052631578947e-05,
"loss": 0.5197,
"step": 70
},
{
"epoch": 35.0,
"eval_accuracy_safe": 0.05428228588570572,
"eval_accuracy_unlabeled": NaN,
"eval_accuracy_unsafe": 0.9955749423043395,
"eval_iou_safe": 0.04808982791841074,
"eval_iou_unlabeled": 0.0,
"eval_iou_unsafe": 0.9655911800769321,
"eval_loss": 0.45792385935783386,
"eval_mean_accuracy": 0.5249286140950227,
"eval_mean_iou": 0.33789366933178094,
"eval_overall_accuracy": 0.9656505584716797,
"eval_runtime": 1.9378,
"eval_samples_per_second": 4.128,
"eval_steps_per_second": 0.516,
"step": 70
},
{
"epoch": 35.5,
"learning_rate": 3.3947368421052636e-05,
"loss": 0.4697,
"step": 71
},
{
"epoch": 36.0,
"learning_rate": 3.368421052631579e-05,
"loss": 0.5519,
"step": 72
},
{
"epoch": 36.5,
"learning_rate": 3.342105263157895e-05,
"loss": 0.4762,
"step": 73
},
{
"epoch": 37.0,
"learning_rate": 3.3157894736842106e-05,
"loss": 0.4986,
"step": 74
},
{
"epoch": 37.5,
"learning_rate": 3.289473684210527e-05,
"loss": 0.4565,
"step": 75
},
{
"epoch": 38.0,
"learning_rate": 3.2631578947368426e-05,
"loss": 0.4981,
"step": 76
},
{
"epoch": 38.5,
"learning_rate": 3.236842105263158e-05,
"loss": 0.4549,
"step": 77
},
{
"epoch": 39.0,
"learning_rate": 3.210526315789474e-05,
"loss": 0.4852,
"step": 78
},
{
"epoch": 39.5,
"learning_rate": 3.1842105263157895e-05,
"loss": 0.4397,
"step": 79
},
{
"epoch": 40.0,
"learning_rate": 3.157894736842105e-05,
"loss": 0.4477,
"step": 80
},
{
"epoch": 40.0,
"eval_accuracy_safe": 0.05825708714564272,
"eval_accuracy_unlabeled": NaN,
"eval_accuracy_unsafe": 0.995995532095335,
"eval_iou_safe": 0.05227315550052489,
"eval_iou_unlabeled": 0.0,
"eval_iou_unsafe": 0.9661213948715597,
"eval_loss": 0.43396273255348206,
"eval_mean_accuracy": 0.5271263096204889,
"eval_mean_iou": 0.3394648501240282,
"eval_overall_accuracy": 0.966184139251709,
"eval_runtime": 1.8717,
"eval_samples_per_second": 4.274,
"eval_steps_per_second": 0.534,
"step": 80
},
{
"epoch": 40.5,
"learning_rate": 3.131578947368421e-05,
"loss": 0.4414,
"step": 81
},
{
"epoch": 41.0,
"learning_rate": 3.105263157894737e-05,
"loss": 0.4533,
"step": 82
},
{
"epoch": 41.5,
"learning_rate": 3.078947368421053e-05,
"loss": 0.4677,
"step": 83
},
{
"epoch": 42.0,
"learning_rate": 3.0526315789473684e-05,
"loss": 0.4149,
"step": 84
},
{
"epoch": 42.5,
"learning_rate": 3.0263157894736844e-05,
"loss": 0.4291,
"step": 85
},
{
"epoch": 43.0,
"learning_rate": 3e-05,
"loss": 0.4353,
"step": 86
},
{
"epoch": 43.5,
"learning_rate": 2.9736842105263157e-05,
"loss": 0.4321,
"step": 87
},
{
"epoch": 44.0,
"learning_rate": 2.9473684210526314e-05,
"loss": 0.4604,
"step": 88
},
{
"epoch": 44.5,
"learning_rate": 2.9210526315789477e-05,
"loss": 0.4617,
"step": 89
},
{
"epoch": 45.0,
"learning_rate": 2.8947368421052634e-05,
"loss": 0.4371,
"step": 90
},
{
"epoch": 45.0,
"eval_accuracy_safe": 0.06065696715164242,
"eval_accuracy_unlabeled": NaN,
"eval_accuracy_unsafe": 0.9966879785193861,
"eval_iou_safe": 0.05525269500348404,
"eval_iou_unlabeled": 0.0,
"eval_iou_unsafe": 0.9668669748527071,
"eval_loss": 0.40330615639686584,
"eval_mean_accuracy": 0.5286724728355142,
"eval_mean_iou": 0.3407065566187304,
"eval_overall_accuracy": 0.9669308662414551,
"eval_runtime": 1.8891,
"eval_samples_per_second": 4.235,
"eval_steps_per_second": 0.529,
"step": 90
},
{
"epoch": 45.5,
"learning_rate": 2.868421052631579e-05,
"loss": 0.422,
"step": 91
},
{
"epoch": 46.0,
"learning_rate": 2.842105263157895e-05,
"loss": 0.4449,
"step": 92
},
{
"epoch": 46.5,
"learning_rate": 2.8157894736842106e-05,
"loss": 0.4311,
"step": 93
},
{
"epoch": 47.0,
"learning_rate": 2.7894736842105263e-05,
"loss": 0.4234,
"step": 94
},
{
"epoch": 47.5,
"learning_rate": 2.7631578947368426e-05,
"loss": 0.4182,
"step": 95
},
{
"epoch": 48.0,
"learning_rate": 2.7368421052631583e-05,
"loss": 0.3954,
"step": 96
},
{
"epoch": 48.5,
"learning_rate": 2.710526315789474e-05,
"loss": 0.4141,
"step": 97
},
{
"epoch": 49.0,
"learning_rate": 2.6842105263157896e-05,
"loss": 0.4147,
"step": 98
},
{
"epoch": 49.5,
"learning_rate": 2.6578947368421052e-05,
"loss": 0.4221,
"step": 99
},
{
"epoch": 50.0,
"learning_rate": 2.6315789473684212e-05,
"loss": 0.3972,
"step": 100
},
{
"epoch": 50.0,
"eval_accuracy_safe": 0.05995200239988001,
"eval_accuracy_unlabeled": NaN,
"eval_accuracy_unsafe": 0.9984525841647451,
"eval_iou_safe": 0.057366343738787226,
"eval_iou_unlabeled": 0.0,
"eval_iou_unsafe": 0.9685570347155371,
"eval_loss": 0.397531658411026,
"eval_mean_accuracy": 0.5292022932823126,
"eval_mean_iou": 0.34197445948477473,
"eval_overall_accuracy": 0.9686169624328613,
"eval_runtime": 1.9669,
"eval_samples_per_second": 4.067,
"eval_steps_per_second": 0.508,
"step": 100
},
{
"epoch": 50.5,
"learning_rate": 2.605263157894737e-05,
"loss": 0.3999,
"step": 101
},
{
"epoch": 51.0,
"learning_rate": 2.578947368421053e-05,
"loss": 0.3632,
"step": 102
},
{
"epoch": 51.5,
"learning_rate": 2.5526315789473688e-05,
"loss": 0.3918,
"step": 103
},
{
"epoch": 52.0,
"learning_rate": 2.5263157894736845e-05,
"loss": 0.4088,
"step": 104
},
{
"epoch": 52.5,
"learning_rate": 2.5e-05,
"loss": 0.3893,
"step": 105
},
{
"epoch": 53.0,
"learning_rate": 2.4736842105263158e-05,
"loss": 0.418,
"step": 106
},
{
"epoch": 53.5,
"learning_rate": 2.4473684210526318e-05,
"loss": 0.3851,
"step": 107
},
{
"epoch": 54.0,
"learning_rate": 2.4210526315789474e-05,
"loss": 0.397,
"step": 108
},
{
"epoch": 54.5,
"learning_rate": 2.394736842105263e-05,
"loss": 0.3817,
"step": 109
},
{
"epoch": 55.0,
"learning_rate": 2.368421052631579e-05,
"loss": 0.4101,
"step": 110
},
{
"epoch": 55.0,
"eval_accuracy_safe": 0.07778611069446528,
"eval_accuracy_unlabeled": NaN,
"eval_accuracy_unsafe": 0.9983156708604164,
"eval_iou_safe": 0.07399058353545442,
"eval_iou_unlabeled": NaN,
"eval_iou_unsafe": 0.9689746391671757,
"eval_loss": 0.37768080830574036,
"eval_mean_accuracy": 0.5380508907774408,
"eval_mean_iou": 0.5214826113513151,
"eval_overall_accuracy": 0.9690513610839844,
"eval_runtime": 1.9039,
"eval_samples_per_second": 4.202,
"eval_steps_per_second": 0.525,
"step": 110
},
{
"epoch": 55.5,
"learning_rate": 2.342105263157895e-05,
"loss": 0.3884,
"step": 111
},
{
"epoch": 56.0,
"learning_rate": 2.3157894736842107e-05,
"loss": 0.4209,
"step": 112
},
{
"epoch": 56.5,
"learning_rate": 2.2894736842105263e-05,
"loss": 0.399,
"step": 113
},
{
"epoch": 57.0,
"learning_rate": 2.2631578947368423e-05,
"loss": 0.3777,
"step": 114
},
{
"epoch": 57.5,
"learning_rate": 2.236842105263158e-05,
"loss": 0.3726,
"step": 115
},
{
"epoch": 58.0,
"learning_rate": 2.2105263157894736e-05,
"loss": 0.3957,
"step": 116
},
{
"epoch": 58.5,
"learning_rate": 2.1842105263157896e-05,
"loss": 0.3829,
"step": 117
},
{
"epoch": 59.0,
"learning_rate": 2.1578947368421053e-05,
"loss": 0.4339,
"step": 118
},
{
"epoch": 59.5,
"learning_rate": 2.1315789473684212e-05,
"loss": 0.371,
"step": 119
},
{
"epoch": 60.0,
"learning_rate": 2.105263157894737e-05,
"loss": 0.3528,
"step": 120
},
{
"epoch": 60.0,
"eval_accuracy_safe": 0.12287385630718464,
"eval_accuracy_unlabeled": NaN,
"eval_accuracy_unsafe": 0.9945485850157746,
"eval_iou_safe": 0.10537825287178894,
"eval_iou_unlabeled": NaN,
"eval_iou_unsafe": 0.9667073567708333,
"eval_loss": 0.3625328540802002,
"eval_mean_accuracy": 0.5587112206614796,
"eval_mean_iou": 0.5360428048213112,
"eval_overall_accuracy": 0.9668374061584473,
"eval_runtime": 1.999,
"eval_samples_per_second": 4.002,
"eval_steps_per_second": 0.5,
"step": 120
}
],
"max_steps": 200,
"num_train_epochs": 100,
"total_flos": 4.21232746954752e+16,
"trial_name": null,
"trial_params": null
}