sam1120's picture
Training in progress, step 120
a8d6991
{
"best_metric": 0.34794068336486816,
"best_model_checkpoint": "/robodata/smodak/Projects/nspl/scripts/terrainseg/training/models/dropoff-utcustom-train-SF-RGBD-b5_4/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": 5.833333333333333e-07,
"loss": 1.0694,
"step": 1
},
{
"epoch": 1.0,
"learning_rate": 1.1666666666666666e-06,
"loss": 1.0723,
"step": 2
},
{
"epoch": 1.5,
"learning_rate": 1.75e-06,
"loss": 1.0706,
"step": 3
},
{
"epoch": 2.0,
"learning_rate": 2.333333333333333e-06,
"loss": 1.065,
"step": 4
},
{
"epoch": 2.5,
"learning_rate": 2.9166666666666666e-06,
"loss": 1.0545,
"step": 5
},
{
"epoch": 3.0,
"learning_rate": 3.5e-06,
"loss": 1.0534,
"step": 6
},
{
"epoch": 3.5,
"learning_rate": 4.083333333333334e-06,
"loss": 1.0415,
"step": 7
},
{
"epoch": 4.0,
"learning_rate": 4.666666666666666e-06,
"loss": 1.0317,
"step": 8
},
{
"epoch": 4.5,
"learning_rate": 5.25e-06,
"loss": 1.0243,
"step": 9
},
{
"epoch": 5.0,
"learning_rate": 5.833333333333333e-06,
"loss": 1.0114,
"step": 10
},
{
"epoch": 5.0,
"eval_accuracy_dropoff": 0.13676001743159247,
"eval_accuracy_undropoff": 0.732156656321488,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.028638451200897586,
"eval_iou_undropoff": 0.7089184137505781,
"eval_iou_unlabeled": 0.0,
"eval_loss": 1.0037065744400024,
"eval_mean_accuracy": 0.43445833687654023,
"eval_mean_iou": 0.24585228831715855,
"eval_overall_accuracy": 0.7074005126953125,
"eval_runtime": 3.5408,
"eval_samples_per_second": 5.648,
"eval_steps_per_second": 0.565,
"step": 10
},
{
"epoch": 5.5,
"learning_rate": 6.4166666666666665e-06,
"loss": 1.0052,
"step": 11
},
{
"epoch": 6.0,
"learning_rate": 7e-06,
"loss": 0.9933,
"step": 12
},
{
"epoch": 6.5,
"learning_rate": 6.969298245614035e-06,
"loss": 0.9842,
"step": 13
},
{
"epoch": 7.0,
"learning_rate": 6.9385964912280696e-06,
"loss": 0.9771,
"step": 14
},
{
"epoch": 7.5,
"learning_rate": 6.907894736842105e-06,
"loss": 0.9683,
"step": 15
},
{
"epoch": 8.0,
"learning_rate": 6.87719298245614e-06,
"loss": 0.9541,
"step": 16
},
{
"epoch": 8.5,
"learning_rate": 6.846491228070175e-06,
"loss": 0.9465,
"step": 17
},
{
"epoch": 9.0,
"learning_rate": 6.815789473684211e-06,
"loss": 0.9306,
"step": 18
},
{
"epoch": 9.5,
"learning_rate": 6.785087719298246e-06,
"loss": 0.9217,
"step": 19
},
{
"epoch": 10.0,
"learning_rate": 6.75438596491228e-06,
"loss": 0.9088,
"step": 20
},
{
"epoch": 10.0,
"eval_accuracy_dropoff": 0.08571297506823551,
"eval_accuracy_undropoff": 0.9235142694807941,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.04595018640749958,
"eval_iou_undropoff": 0.8896697494651115,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.8245114088058472,
"eval_mean_accuracy": 0.5046136222745148,
"eval_mean_iou": 0.31187331195753704,
"eval_overall_accuracy": 0.8886791229248047,
"eval_runtime": 2.8775,
"eval_samples_per_second": 6.95,
"eval_steps_per_second": 0.695,
"step": 20
},
{
"epoch": 10.5,
"learning_rate": 6.7236842105263156e-06,
"loss": 0.9041,
"step": 21
},
{
"epoch": 11.0,
"learning_rate": 6.692982456140351e-06,
"loss": 0.877,
"step": 22
},
{
"epoch": 11.5,
"learning_rate": 6.662280701754385e-06,
"loss": 0.8683,
"step": 23
},
{
"epoch": 12.0,
"learning_rate": 6.6315789473684205e-06,
"loss": 0.8695,
"step": 24
},
{
"epoch": 12.5,
"learning_rate": 6.600877192982457e-06,
"loss": 0.8496,
"step": 25
},
{
"epoch": 13.0,
"learning_rate": 6.570175438596491e-06,
"loss": 0.8368,
"step": 26
},
{
"epoch": 13.5,
"learning_rate": 6.539473684210526e-06,
"loss": 0.825,
"step": 27
},
{
"epoch": 14.0,
"learning_rate": 6.5087719298245616e-06,
"loss": 0.8255,
"step": 28
},
{
"epoch": 14.5,
"learning_rate": 6.478070175438596e-06,
"loss": 0.8017,
"step": 29
},
{
"epoch": 15.0,
"learning_rate": 6.447368421052631e-06,
"loss": 0.8029,
"step": 30
},
{
"epoch": 15.0,
"eval_accuracy_dropoff": 0.03994587031812656,
"eval_accuracy_undropoff": 0.9596368872123442,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.02525287676317743,
"eval_iou_undropoff": 0.921926919826706,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.6619901061058044,
"eval_mean_accuracy": 0.4997913787652354,
"eval_mean_iou": 0.3157265988632945,
"eval_overall_accuracy": 0.9213968276977539,
"eval_runtime": 2.9201,
"eval_samples_per_second": 6.849,
"eval_steps_per_second": 0.685,
"step": 30
},
{
"epoch": 15.5,
"learning_rate": 6.4166666666666665e-06,
"loss": 0.7827,
"step": 31
},
{
"epoch": 16.0,
"learning_rate": 6.385964912280701e-06,
"loss": 0.7741,
"step": 32
},
{
"epoch": 16.5,
"learning_rate": 6.355263157894737e-06,
"loss": 0.7773,
"step": 33
},
{
"epoch": 17.0,
"learning_rate": 6.324561403508772e-06,
"loss": 0.7609,
"step": 34
},
{
"epoch": 17.5,
"learning_rate": 6.293859649122807e-06,
"loss": 0.7477,
"step": 35
},
{
"epoch": 18.0,
"learning_rate": 6.263157894736842e-06,
"loss": 0.7405,
"step": 36
},
{
"epoch": 18.5,
"learning_rate": 6.232456140350877e-06,
"loss": 0.7221,
"step": 37
},
{
"epoch": 19.0,
"learning_rate": 6.201754385964912e-06,
"loss": 0.7194,
"step": 38
},
{
"epoch": 19.5,
"learning_rate": 6.171052631578947e-06,
"loss": 0.7267,
"step": 39
},
{
"epoch": 20.0,
"learning_rate": 6.140350877192982e-06,
"loss": 0.6935,
"step": 40
},
{
"epoch": 20.0,
"eval_accuracy_dropoff": 0.021431684212940665,
"eval_accuracy_undropoff": 0.9703820087424886,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.015123901396824369,
"eval_iou_undropoff": 0.9311250629685884,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.5661563873291016,
"eval_mean_accuracy": 0.4959068464777146,
"eval_mean_iou": 0.3154163214551376,
"eval_overall_accuracy": 0.9309253692626953,
"eval_runtime": 2.9118,
"eval_samples_per_second": 6.869,
"eval_steps_per_second": 0.687,
"step": 40
},
{
"epoch": 20.5,
"learning_rate": 6.1096491228070174e-06,
"loss": 0.691,
"step": 41
},
{
"epoch": 21.0,
"learning_rate": 6.078947368421053e-06,
"loss": 0.6723,
"step": 42
},
{
"epoch": 21.5,
"learning_rate": 6.048245614035088e-06,
"loss": 0.6811,
"step": 43
},
{
"epoch": 22.0,
"learning_rate": 6.017543859649122e-06,
"loss": 0.6487,
"step": 44
},
{
"epoch": 22.5,
"learning_rate": 5.986842105263158e-06,
"loss": 0.6643,
"step": 45
},
{
"epoch": 23.0,
"learning_rate": 5.956140350877193e-06,
"loss": 0.6587,
"step": 46
},
{
"epoch": 23.5,
"learning_rate": 5.925438596491228e-06,
"loss": 0.6371,
"step": 47
},
{
"epoch": 24.0,
"learning_rate": 5.894736842105263e-06,
"loss": 0.6436,
"step": 48
},
{
"epoch": 24.5,
"learning_rate": 5.864035087719299e-06,
"loss": 0.6312,
"step": 49
},
{
"epoch": 25.0,
"learning_rate": 5.833333333333333e-06,
"loss": 0.635,
"step": 50
},
{
"epoch": 25.0,
"eval_accuracy_dropoff": 0.01526640519277965,
"eval_accuracy_undropoff": 0.9802598467427612,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.012095308706587,
"eval_iou_undropoff": 0.9403544093266979,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.5018354058265686,
"eval_mean_accuracy": 0.49776312596777045,
"eval_mean_iou": 0.3174832393444283,
"eval_overall_accuracy": 0.9401361465454101,
"eval_runtime": 13.2377,
"eval_samples_per_second": 1.511,
"eval_steps_per_second": 0.151,
"step": 50
},
{
"epoch": 25.5,
"learning_rate": 5.802631578947368e-06,
"loss": 0.6097,
"step": 51
},
{
"epoch": 26.0,
"learning_rate": 5.771929824561404e-06,
"loss": 0.6166,
"step": 52
},
{
"epoch": 26.5,
"learning_rate": 5.741228070175439e-06,
"loss": 0.6093,
"step": 53
},
{
"epoch": 27.0,
"learning_rate": 5.710526315789473e-06,
"loss": 0.6055,
"step": 54
},
{
"epoch": 27.5,
"learning_rate": 5.6798245614035086e-06,
"loss": 0.5944,
"step": 55
},
{
"epoch": 28.0,
"learning_rate": 5.649122807017544e-06,
"loss": 0.574,
"step": 56
},
{
"epoch": 28.5,
"learning_rate": 5.618421052631579e-06,
"loss": 0.5909,
"step": 57
},
{
"epoch": 29.0,
"learning_rate": 5.587719298245614e-06,
"loss": 0.5751,
"step": 58
},
{
"epoch": 29.5,
"learning_rate": 5.55701754385965e-06,
"loss": 0.5794,
"step": 59
},
{
"epoch": 30.0,
"learning_rate": 5.526315789473684e-06,
"loss": 0.5579,
"step": 60
},
{
"epoch": 30.0,
"eval_accuracy_dropoff": 0.013050758044909287,
"eval_accuracy_undropoff": 0.9824573099682878,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.011141788560574909,
"eval_iou_undropoff": 0.9423118438942378,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.4701308310031891,
"eval_mean_accuracy": 0.49775403400659857,
"eval_mean_iou": 0.31781787748493756,
"eval_overall_accuracy": 0.9421501159667969,
"eval_runtime": 2.835,
"eval_samples_per_second": 7.055,
"eval_steps_per_second": 0.705,
"step": 60
},
{
"epoch": 30.5,
"learning_rate": 5.495614035087719e-06,
"loss": 0.5654,
"step": 61
},
{
"epoch": 31.0,
"learning_rate": 5.4649122807017546e-06,
"loss": 0.5423,
"step": 62
},
{
"epoch": 31.5,
"learning_rate": 5.434210526315789e-06,
"loss": 0.5493,
"step": 63
},
{
"epoch": 32.0,
"learning_rate": 5.403508771929824e-06,
"loss": 0.5479,
"step": 64
},
{
"epoch": 32.5,
"learning_rate": 5.37280701754386e-06,
"loss": 0.5559,
"step": 65
},
{
"epoch": 33.0,
"learning_rate": 5.342105263157895e-06,
"loss": 0.5416,
"step": 66
},
{
"epoch": 33.5,
"learning_rate": 5.31140350877193e-06,
"loss": 0.5307,
"step": 67
},
{
"epoch": 34.0,
"learning_rate": 5.280701754385965e-06,
"loss": 0.5303,
"step": 68
},
{
"epoch": 34.5,
"learning_rate": 5.25e-06,
"loss": 0.5238,
"step": 69
},
{
"epoch": 35.0,
"learning_rate": 5.219298245614035e-06,
"loss": 0.5086,
"step": 70
},
{
"epoch": 35.0,
"eval_accuracy_dropoff": 0.008844239546778597,
"eval_accuracy_undropoff": 0.9865837725639492,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.008049062543052883,
"eval_iou_undropoff": 0.9461391183132418,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.44029974937438965,
"eval_mean_accuracy": 0.4977140060553639,
"eval_mean_iou": 0.31806272695209825,
"eval_overall_accuracy": 0.9459300994873047,
"eval_runtime": 2.9283,
"eval_samples_per_second": 6.83,
"eval_steps_per_second": 0.683,
"step": 70
},
{
"epoch": 35.5,
"learning_rate": 5.18859649122807e-06,
"loss": 0.5131,
"step": 71
},
{
"epoch": 36.0,
"learning_rate": 5.157894736842105e-06,
"loss": 0.5041,
"step": 72
},
{
"epoch": 36.5,
"learning_rate": 5.12719298245614e-06,
"loss": 0.528,
"step": 73
},
{
"epoch": 37.0,
"learning_rate": 5.096491228070176e-06,
"loss": 0.4949,
"step": 74
},
{
"epoch": 37.5,
"learning_rate": 5.0657894736842104e-06,
"loss": 0.498,
"step": 75
},
{
"epoch": 38.0,
"learning_rate": 5.035087719298246e-06,
"loss": 0.5077,
"step": 76
},
{
"epoch": 38.5,
"learning_rate": 5.004385964912281e-06,
"loss": 0.4977,
"step": 77
},
{
"epoch": 39.0,
"learning_rate": 4.973684210526315e-06,
"loss": 0.4885,
"step": 78
},
{
"epoch": 39.5,
"learning_rate": 4.942982456140351e-06,
"loss": 0.4882,
"step": 79
},
{
"epoch": 40.0,
"learning_rate": 4.912280701754386e-06,
"loss": 0.472,
"step": 80
},
{
"epoch": 40.0,
"eval_accuracy_dropoff": 0.006284547810729604,
"eval_accuracy_undropoff": 0.9878638018581519,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.005870480908081981,
"eval_iou_undropoff": 0.9472967064759237,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.43275266885757446,
"eval_mean_accuracy": 0.4970741748344408,
"eval_mean_iou": 0.31772239579466854,
"eval_overall_accuracy": 0.9470504760742188,
"eval_runtime": 2.9365,
"eval_samples_per_second": 6.811,
"eval_steps_per_second": 0.681,
"step": 80
},
{
"epoch": 40.5,
"learning_rate": 4.88157894736842e-06,
"loss": 0.4733,
"step": 81
},
{
"epoch": 41.0,
"learning_rate": 4.8508771929824564e-06,
"loss": 0.4665,
"step": 82
},
{
"epoch": 41.5,
"learning_rate": 4.820175438596492e-06,
"loss": 0.4608,
"step": 83
},
{
"epoch": 42.0,
"learning_rate": 4.789473684210526e-06,
"loss": 0.4715,
"step": 84
},
{
"epoch": 42.5,
"learning_rate": 4.758771929824561e-06,
"loss": 0.4654,
"step": 85
},
{
"epoch": 43.0,
"learning_rate": 4.728070175438597e-06,
"loss": 0.45,
"step": 86
},
{
"epoch": 43.5,
"learning_rate": 4.697368421052631e-06,
"loss": 0.4555,
"step": 87
},
{
"epoch": 44.0,
"learning_rate": 4.666666666666666e-06,
"loss": 0.4494,
"step": 88
},
{
"epoch": 44.5,
"learning_rate": 4.635964912280702e-06,
"loss": 0.4599,
"step": 89
},
{
"epoch": 45.0,
"learning_rate": 4.605263157894737e-06,
"loss": 0.4484,
"step": 90
},
{
"epoch": 45.0,
"eval_accuracy_dropoff": 0.004587261175715039,
"eval_accuracy_undropoff": 0.991641798767534,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.004395044148218469,
"eval_iou_undropoff": 0.9507666029246044,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.41358834505081177,
"eval_mean_accuracy": 0.49811452997162453,
"eval_mean_iou": 0.3183872156909409,
"eval_overall_accuracy": 0.9506008148193359,
"eval_runtime": 2.9507,
"eval_samples_per_second": 6.778,
"eval_steps_per_second": 0.678,
"step": 90
},
{
"epoch": 45.5,
"learning_rate": 4.574561403508772e-06,
"loss": 0.4507,
"step": 91
},
{
"epoch": 46.0,
"learning_rate": 4.543859649122807e-06,
"loss": 0.4315,
"step": 92
},
{
"epoch": 46.5,
"learning_rate": 4.513157894736842e-06,
"loss": 0.4204,
"step": 93
},
{
"epoch": 47.0,
"learning_rate": 4.482456140350877e-06,
"loss": 0.4494,
"step": 94
},
{
"epoch": 47.5,
"learning_rate": 4.451754385964912e-06,
"loss": 0.4303,
"step": 95
},
{
"epoch": 48.0,
"learning_rate": 4.421052631578947e-06,
"loss": 0.4353,
"step": 96
},
{
"epoch": 48.5,
"learning_rate": 4.390350877192982e-06,
"loss": 0.434,
"step": 97
},
{
"epoch": 49.0,
"learning_rate": 4.359649122807018e-06,
"loss": 0.4266,
"step": 98
},
{
"epoch": 49.5,
"learning_rate": 4.3289473684210525e-06,
"loss": 0.4177,
"step": 99
},
{
"epoch": 50.0,
"learning_rate": 4.298245614035088e-06,
"loss": 0.4026,
"step": 100
},
{
"epoch": 50.0,
"eval_accuracy_dropoff": 0.004348723594577857,
"eval_accuracy_undropoff": 0.9926461998632804,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.004242521883894528,
"eval_iou_undropoff": 0.9516593303838112,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.40131449699401855,
"eval_mean_accuracy": 0.4984974617289291,
"eval_mean_iou": 0.3186339507559019,
"eval_overall_accuracy": 0.9515535354614257,
"eval_runtime": 5.77,
"eval_samples_per_second": 3.466,
"eval_steps_per_second": 0.347,
"step": 100
},
{
"epoch": 50.5,
"learning_rate": 4.267543859649123e-06,
"loss": 0.3999,
"step": 101
},
{
"epoch": 51.0,
"learning_rate": 4.2368421052631575e-06,
"loss": 0.4064,
"step": 102
},
{
"epoch": 51.5,
"learning_rate": 4.206140350877193e-06,
"loss": 0.413,
"step": 103
},
{
"epoch": 52.0,
"learning_rate": 4.175438596491228e-06,
"loss": 0.403,
"step": 104
},
{
"epoch": 52.5,
"learning_rate": 4.144736842105262e-06,
"loss": 0.3948,
"step": 105
},
{
"epoch": 53.0,
"learning_rate": 4.1140350877192985e-06,
"loss": 0.3892,
"step": 106
},
{
"epoch": 53.5,
"learning_rate": 4.083333333333334e-06,
"loss": 0.3852,
"step": 107
},
{
"epoch": 54.0,
"learning_rate": 4.052631578947368e-06,
"loss": 0.3905,
"step": 108
},
{
"epoch": 54.5,
"learning_rate": 4.0219298245614034e-06,
"loss": 0.386,
"step": 109
},
{
"epoch": 55.0,
"learning_rate": 3.991228070175439e-06,
"loss": 0.3873,
"step": 110
},
{
"epoch": 55.0,
"eval_accuracy_dropoff": 0.0009587375857244432,
"eval_accuracy_undropoff": 0.9971428201839445,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.0009477297564470564,
"eval_iou_undropoff": 0.9557326855127719,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.36210715770721436,
"eval_mean_accuracy": 0.49905077888483446,
"eval_mean_iou": 0.3188934717564063,
"eval_overall_accuracy": 0.9557222366333008,
"eval_runtime": 2.9583,
"eval_samples_per_second": 6.761,
"eval_steps_per_second": 0.676,
"step": 110
},
{
"epoch": 55.5,
"learning_rate": 3.960526315789473e-06,
"loss": 0.3824,
"step": 111
},
{
"epoch": 56.0,
"learning_rate": 3.929824561403508e-06,
"loss": 0.3976,
"step": 112
},
{
"epoch": 56.5,
"learning_rate": 3.899122807017544e-06,
"loss": 0.3765,
"step": 113
},
{
"epoch": 57.0,
"learning_rate": 3.868421052631579e-06,
"loss": 0.3774,
"step": 114
},
{
"epoch": 57.5,
"learning_rate": 3.837719298245614e-06,
"loss": 0.3994,
"step": 115
},
{
"epoch": 58.0,
"learning_rate": 3.8070175438596494e-06,
"loss": 0.3556,
"step": 116
},
{
"epoch": 58.5,
"learning_rate": 3.7763157894736843e-06,
"loss": 0.3692,
"step": 117
},
{
"epoch": 59.0,
"learning_rate": 3.745614035087719e-06,
"loss": 0.3648,
"step": 118
},
{
"epoch": 59.5,
"learning_rate": 3.7149122807017544e-06,
"loss": 0.3674,
"step": 119
},
{
"epoch": 60.0,
"learning_rate": 3.6842105263157892e-06,
"loss": 0.3549,
"step": 120
},
{
"epoch": 60.0,
"eval_accuracy_dropoff": 0.00042202802816578364,
"eval_accuracy_undropoff": 0.9978805485100654,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.0004183586698013251,
"eval_iou_undropoff": 0.9564140540880779,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.34794068336486816,
"eval_mean_accuracy": 0.4991512882691156,
"eval_mean_iou": 0.31894413758595974,
"eval_overall_accuracy": 0.9564069747924805,
"eval_runtime": 2.8713,
"eval_samples_per_second": 6.965,
"eval_steps_per_second": 0.697,
"step": 120
}
],
"max_steps": 240,
"num_train_epochs": 120,
"total_flos": 1.3892743058581094e+18,
"trial_name": null,
"trial_params": null
}