sam1120's picture
Training in progress, step 120
922ff31
raw
history blame
22 kB
{
"best_metric": 0.2761629819869995,
"best_model_checkpoint": "/robodata/smodak/Projects/nspl/scripts/terrainseg/training/models/dropoff-utcustom-train-SF-RGBD-b0_7/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": 6.666666666666667e-06,
"loss": 1.0264,
"step": 1
},
{
"epoch": 1.0,
"learning_rate": 1.3333333333333333e-05,
"loss": 1.0249,
"step": 2
},
{
"epoch": 1.5,
"learning_rate": 2e-05,
"loss": 1.0219,
"step": 3
},
{
"epoch": 2.0,
"learning_rate": 2.6666666666666667e-05,
"loss": 1.0261,
"step": 4
},
{
"epoch": 2.5,
"learning_rate": 3.3333333333333335e-05,
"loss": 1.0043,
"step": 5
},
{
"epoch": 3.0,
"learning_rate": 4e-05,
"loss": 0.9882,
"step": 6
},
{
"epoch": 3.5,
"learning_rate": 4.666666666666667e-05,
"loss": 0.9908,
"step": 7
},
{
"epoch": 4.0,
"learning_rate": 5.333333333333333e-05,
"loss": 0.9709,
"step": 8
},
{
"epoch": 4.5,
"learning_rate": 6.000000000000001e-05,
"loss": 0.961,
"step": 9
},
{
"epoch": 5.0,
"learning_rate": 6.666666666666667e-05,
"loss": 0.9508,
"step": 10
},
{
"epoch": 5.0,
"eval_accuracy_dropoff": 0.19369710314456753,
"eval_accuracy_undropoff": 0.9010801242217483,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.06049444267748618,
"eval_iou_undropoff": 0.8706257192603624,
"eval_iou_unlabeled": 0.0,
"eval_loss": 1.026282548904419,
"eval_mean_accuracy": 0.5473886136831579,
"eval_mean_iou": 0.31037338731261616,
"eval_overall_accuracy": 0.8716676712036133,
"eval_runtime": 2.7665,
"eval_samples_per_second": 7.229,
"eval_steps_per_second": 0.723,
"step": 10
},
{
"epoch": 5.5,
"learning_rate": 7.333333333333333e-05,
"loss": 0.9292,
"step": 11
},
{
"epoch": 6.0,
"learning_rate": 8e-05,
"loss": 0.9246,
"step": 12
},
{
"epoch": 6.5,
"learning_rate": 7.964912280701755e-05,
"loss": 0.9097,
"step": 13
},
{
"epoch": 7.0,
"learning_rate": 7.929824561403509e-05,
"loss": 0.8786,
"step": 14
},
{
"epoch": 7.5,
"learning_rate": 7.894736842105263e-05,
"loss": 0.8697,
"step": 15
},
{
"epoch": 8.0,
"learning_rate": 7.859649122807017e-05,
"loss": 0.8421,
"step": 16
},
{
"epoch": 8.5,
"learning_rate": 7.824561403508773e-05,
"loss": 0.8223,
"step": 17
},
{
"epoch": 9.0,
"learning_rate": 7.789473684210527e-05,
"loss": 0.8166,
"step": 18
},
{
"epoch": 9.5,
"learning_rate": 7.754385964912281e-05,
"loss": 0.7987,
"step": 19
},
{
"epoch": 10.0,
"learning_rate": 7.719298245614036e-05,
"loss": 0.7814,
"step": 20
},
{
"epoch": 10.0,
"eval_accuracy_dropoff": 0.09518566939608707,
"eval_accuracy_undropoff": 0.9726282293027602,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.05836274909643214,
"eval_iou_undropoff": 0.9358911019066932,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.7567641139030457,
"eval_mean_accuracy": 0.5339069493494236,
"eval_mean_iou": 0.4971269255015627,
"eval_overall_accuracy": 0.9361448287963867,
"eval_runtime": 2.8406,
"eval_samples_per_second": 7.041,
"eval_steps_per_second": 0.704,
"step": 20
},
{
"epoch": 10.5,
"learning_rate": 7.68421052631579e-05,
"loss": 0.7775,
"step": 21
},
{
"epoch": 11.0,
"learning_rate": 7.649122807017544e-05,
"loss": 0.744,
"step": 22
},
{
"epoch": 11.5,
"learning_rate": 7.614035087719298e-05,
"loss": 0.7385,
"step": 23
},
{
"epoch": 12.0,
"learning_rate": 7.578947368421052e-05,
"loss": 0.7156,
"step": 24
},
{
"epoch": 12.5,
"learning_rate": 7.543859649122808e-05,
"loss": 0.7073,
"step": 25
},
{
"epoch": 13.0,
"learning_rate": 7.508771929824562e-05,
"loss": 0.6876,
"step": 26
},
{
"epoch": 13.5,
"learning_rate": 7.473684210526316e-05,
"loss": 0.68,
"step": 27
},
{
"epoch": 14.0,
"learning_rate": 7.43859649122807e-05,
"loss": 0.6647,
"step": 28
},
{
"epoch": 14.5,
"learning_rate": 7.403508771929825e-05,
"loss": 0.656,
"step": 29
},
{
"epoch": 15.0,
"learning_rate": 7.368421052631579e-05,
"loss": 0.642,
"step": 30
},
{
"epoch": 15.0,
"eval_accuracy_dropoff": 0.10256657262781256,
"eval_accuracy_undropoff": 0.9861238615411099,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.07771069890623208,
"eval_iou_undropoff": 0.9491694411343236,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.590660810470581,
"eval_mean_accuracy": 0.5443452170844613,
"eval_mean_iou": 0.5134400700202778,
"eval_overall_accuracy": 0.9493862152099609,
"eval_runtime": 2.7723,
"eval_samples_per_second": 7.214,
"eval_steps_per_second": 0.721,
"step": 30
},
{
"epoch": 15.5,
"learning_rate": 7.333333333333333e-05,
"loss": 0.6219,
"step": 31
},
{
"epoch": 16.0,
"learning_rate": 7.298245614035087e-05,
"loss": 0.6276,
"step": 32
},
{
"epoch": 16.5,
"learning_rate": 7.263157894736843e-05,
"loss": 0.5895,
"step": 33
},
{
"epoch": 17.0,
"learning_rate": 7.228070175438597e-05,
"loss": 0.5987,
"step": 34
},
{
"epoch": 17.5,
"learning_rate": 7.192982456140351e-05,
"loss": 0.578,
"step": 35
},
{
"epoch": 18.0,
"learning_rate": 7.157894736842105e-05,
"loss": 0.5622,
"step": 36
},
{
"epoch": 18.5,
"learning_rate": 7.12280701754386e-05,
"loss": 0.5411,
"step": 37
},
{
"epoch": 19.0,
"learning_rate": 7.087719298245614e-05,
"loss": 0.5512,
"step": 38
},
{
"epoch": 19.5,
"learning_rate": 7.052631578947368e-05,
"loss": 0.5536,
"step": 39
},
{
"epoch": 20.0,
"learning_rate": 7.017543859649124e-05,
"loss": 0.5118,
"step": 40
},
{
"epoch": 20.0,
"eval_accuracy_dropoff": 0.20064680382577582,
"eval_accuracy_undropoff": 0.9839122686389837,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.146396810999508,
"eval_iou_undropoff": 0.9509353469996961,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.48036208748817444,
"eval_mean_accuracy": 0.5922795362323798,
"eval_mean_iou": 0.3657773859997347,
"eval_overall_accuracy": 0.9513446807861328,
"eval_runtime": 3.0452,
"eval_samples_per_second": 6.568,
"eval_steps_per_second": 0.657,
"step": 40
},
{
"epoch": 20.5,
"learning_rate": 6.982456140350878e-05,
"loss": 0.5335,
"step": 41
},
{
"epoch": 21.0,
"learning_rate": 6.947368421052632e-05,
"loss": 0.5013,
"step": 42
},
{
"epoch": 21.5,
"learning_rate": 6.912280701754386e-05,
"loss": 0.538,
"step": 43
},
{
"epoch": 22.0,
"learning_rate": 6.87719298245614e-05,
"loss": 0.5046,
"step": 44
},
{
"epoch": 22.5,
"learning_rate": 6.842105263157895e-05,
"loss": 0.4906,
"step": 45
},
{
"epoch": 23.0,
"learning_rate": 6.80701754385965e-05,
"loss": 0.4797,
"step": 46
},
{
"epoch": 23.5,
"learning_rate": 6.771929824561404e-05,
"loss": 0.492,
"step": 47
},
{
"epoch": 24.0,
"learning_rate": 6.736842105263159e-05,
"loss": 0.4671,
"step": 48
},
{
"epoch": 24.5,
"learning_rate": 6.701754385964913e-05,
"loss": 0.477,
"step": 49
},
{
"epoch": 25.0,
"learning_rate": 6.666666666666667e-05,
"loss": 0.4581,
"step": 50
},
{
"epoch": 25.0,
"eval_accuracy_dropoff": 0.19300442670703458,
"eval_accuracy_undropoff": 0.9900139008156406,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.15784833444007998,
"eval_iou_undropoff": 0.9565325903664265,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.44047123193740845,
"eval_mean_accuracy": 0.5915091637613376,
"eval_mean_iou": 0.37146030826883547,
"eval_overall_accuracy": 0.9568748474121094,
"eval_runtime": 2.862,
"eval_samples_per_second": 6.988,
"eval_steps_per_second": 0.699,
"step": 50
},
{
"epoch": 25.5,
"learning_rate": 6.631578947368421e-05,
"loss": 0.451,
"step": 51
},
{
"epoch": 26.0,
"learning_rate": 6.596491228070175e-05,
"loss": 0.4583,
"step": 52
},
{
"epoch": 26.5,
"learning_rate": 6.561403508771931e-05,
"loss": 0.4437,
"step": 53
},
{
"epoch": 27.0,
"learning_rate": 6.526315789473685e-05,
"loss": 0.4363,
"step": 54
},
{
"epoch": 27.5,
"learning_rate": 6.49122807017544e-05,
"loss": 0.4311,
"step": 55
},
{
"epoch": 28.0,
"learning_rate": 6.456140350877194e-05,
"loss": 0.4346,
"step": 56
},
{
"epoch": 28.5,
"learning_rate": 6.421052631578948e-05,
"loss": 0.419,
"step": 57
},
{
"epoch": 29.0,
"learning_rate": 6.385964912280702e-05,
"loss": 0.4368,
"step": 58
},
{
"epoch": 29.5,
"learning_rate": 6.350877192982457e-05,
"loss": 0.4127,
"step": 59
},
{
"epoch": 30.0,
"learning_rate": 6.315789473684212e-05,
"loss": 0.4213,
"step": 60
},
{
"epoch": 30.0,
"eval_accuracy_dropoff": 0.23790453909493336,
"eval_accuracy_undropoff": 0.9892081112304063,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.19095414478964925,
"eval_iou_undropoff": 0.9575495596432145,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.4145582318305969,
"eval_mean_accuracy": 0.6135563251626699,
"eval_mean_iou": 0.38283456814428796,
"eval_overall_accuracy": 0.9579694747924805,
"eval_runtime": 2.9309,
"eval_samples_per_second": 6.824,
"eval_steps_per_second": 0.682,
"step": 60
},
{
"epoch": 30.5,
"learning_rate": 6.280701754385966e-05,
"loss": 0.4109,
"step": 61
},
{
"epoch": 31.0,
"learning_rate": 6.24561403508772e-05,
"loss": 0.3886,
"step": 62
},
{
"epoch": 31.5,
"learning_rate": 6.210526315789474e-05,
"loss": 0.4031,
"step": 63
},
{
"epoch": 32.0,
"learning_rate": 6.175438596491228e-05,
"loss": 0.3899,
"step": 64
},
{
"epoch": 32.5,
"learning_rate": 6.140350877192984e-05,
"loss": 0.3879,
"step": 65
},
{
"epoch": 33.0,
"learning_rate": 6.105263157894738e-05,
"loss": 0.3708,
"step": 66
},
{
"epoch": 33.5,
"learning_rate": 6.070175438596492e-05,
"loss": 0.382,
"step": 67
},
{
"epoch": 34.0,
"learning_rate": 6.035087719298246e-05,
"loss": 0.3716,
"step": 68
},
{
"epoch": 34.5,
"learning_rate": 6.000000000000001e-05,
"loss": 0.3808,
"step": 69
},
{
"epoch": 35.0,
"learning_rate": 5.964912280701755e-05,
"loss": 0.3571,
"step": 70
},
{
"epoch": 35.0,
"eval_accuracy_dropoff": 0.2473726461616092,
"eval_accuracy_undropoff": 0.9886681983766793,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.19633228600450728,
"eval_iou_undropoff": 0.9574076008382422,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.375016450881958,
"eval_mean_accuracy": 0.6180204222691442,
"eval_mean_iou": 0.3845799622809165,
"eval_overall_accuracy": 0.9578456878662109,
"eval_runtime": 2.8583,
"eval_samples_per_second": 6.997,
"eval_steps_per_second": 0.7,
"step": 70
},
{
"epoch": 35.5,
"learning_rate": 5.929824561403509e-05,
"loss": 0.367,
"step": 71
},
{
"epoch": 36.0,
"learning_rate": 5.8947368421052634e-05,
"loss": 0.3767,
"step": 72
},
{
"epoch": 36.5,
"learning_rate": 5.8596491228070176e-05,
"loss": 0.3713,
"step": 73
},
{
"epoch": 37.0,
"learning_rate": 5.8245614035087725e-05,
"loss": 0.3326,
"step": 74
},
{
"epoch": 37.5,
"learning_rate": 5.789473684210527e-05,
"loss": 0.348,
"step": 75
},
{
"epoch": 38.0,
"learning_rate": 5.7543859649122816e-05,
"loss": 0.3401,
"step": 76
},
{
"epoch": 38.5,
"learning_rate": 5.719298245614036e-05,
"loss": 0.3436,
"step": 77
},
{
"epoch": 39.0,
"learning_rate": 5.68421052631579e-05,
"loss": 0.3487,
"step": 78
},
{
"epoch": 39.5,
"learning_rate": 5.649122807017544e-05,
"loss": 0.3398,
"step": 79
},
{
"epoch": 40.0,
"learning_rate": 5.6140350877192984e-05,
"loss": 0.3205,
"step": 80
},
{
"epoch": 40.0,
"eval_accuracy_dropoff": 0.2522076194408129,
"eval_accuracy_undropoff": 0.9881893814485306,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.19823899733902547,
"eval_iou_undropoff": 0.9571383411399603,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.3478124737739563,
"eval_mean_accuracy": 0.6201985004446717,
"eval_mean_iou": 0.5776886692394929,
"eval_overall_accuracy": 0.9575878143310547,
"eval_runtime": 2.8448,
"eval_samples_per_second": 7.03,
"eval_steps_per_second": 0.703,
"step": 80
},
{
"epoch": 40.5,
"learning_rate": 5.5789473684210526e-05,
"loss": 0.3368,
"step": 81
},
{
"epoch": 41.0,
"learning_rate": 5.543859649122808e-05,
"loss": 0.3154,
"step": 82
},
{
"epoch": 41.5,
"learning_rate": 5.508771929824562e-05,
"loss": 0.3346,
"step": 83
},
{
"epoch": 42.0,
"learning_rate": 5.4736842105263165e-05,
"loss": 0.32,
"step": 84
},
{
"epoch": 42.5,
"learning_rate": 5.438596491228071e-05,
"loss": 0.3161,
"step": 85
},
{
"epoch": 43.0,
"learning_rate": 5.403508771929825e-05,
"loss": 0.3033,
"step": 86
},
{
"epoch": 43.5,
"learning_rate": 5.368421052631579e-05,
"loss": 0.3246,
"step": 87
},
{
"epoch": 44.0,
"learning_rate": 5.333333333333333e-05,
"loss": 0.3147,
"step": 88
},
{
"epoch": 44.5,
"learning_rate": 5.2982456140350875e-05,
"loss": 0.3171,
"step": 89
},
{
"epoch": 45.0,
"learning_rate": 5.263157894736843e-05,
"loss": 0.3114,
"step": 90
},
{
"epoch": 45.0,
"eval_accuracy_dropoff": 0.30220417899493107,
"eval_accuracy_undropoff": 0.9823902437568223,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.21497190442937603,
"eval_iou_undropoff": 0.9535308994629481,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.3460918366909027,
"eval_mean_accuracy": 0.6422972113758767,
"eval_mean_iou": 0.38950093463077473,
"eval_overall_accuracy": 0.9541086196899414,
"eval_runtime": 2.7794,
"eval_samples_per_second": 7.196,
"eval_steps_per_second": 0.72,
"step": 90
},
{
"epoch": 45.5,
"learning_rate": 5.228070175438597e-05,
"loss": 0.3018,
"step": 91
},
{
"epoch": 46.0,
"learning_rate": 5.1929824561403515e-05,
"loss": 0.295,
"step": 92
},
{
"epoch": 46.5,
"learning_rate": 5.157894736842106e-05,
"loss": 0.2985,
"step": 93
},
{
"epoch": 47.0,
"learning_rate": 5.12280701754386e-05,
"loss": 0.2946,
"step": 94
},
{
"epoch": 47.5,
"learning_rate": 5.087719298245614e-05,
"loss": 0.302,
"step": 95
},
{
"epoch": 48.0,
"learning_rate": 5.052631578947368e-05,
"loss": 0.2771,
"step": 96
},
{
"epoch": 48.5,
"learning_rate": 5.017543859649123e-05,
"loss": 0.2849,
"step": 97
},
{
"epoch": 49.0,
"learning_rate": 4.982456140350878e-05,
"loss": 0.2958,
"step": 98
},
{
"epoch": 49.5,
"learning_rate": 4.947368421052632e-05,
"loss": 0.2795,
"step": 99
},
{
"epoch": 50.0,
"learning_rate": 4.9122807017543864e-05,
"loss": 0.2747,
"step": 100
},
{
"epoch": 50.0,
"eval_accuracy_dropoff": 0.28468084130369964,
"eval_accuracy_undropoff": 0.9867370099017192,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.2180262788083193,
"eval_iou_undropoff": 0.9570375042874478,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.3253457546234131,
"eval_mean_accuracy": 0.6357089256027094,
"eval_mean_iou": 0.5875318915478835,
"eval_overall_accuracy": 0.9575460433959961,
"eval_runtime": 2.8813,
"eval_samples_per_second": 6.941,
"eval_steps_per_second": 0.694,
"step": 100
},
{
"epoch": 50.5,
"learning_rate": 4.8771929824561406e-05,
"loss": 0.2742,
"step": 101
},
{
"epoch": 51.0,
"learning_rate": 4.842105263157895e-05,
"loss": 0.2787,
"step": 102
},
{
"epoch": 51.5,
"learning_rate": 4.80701754385965e-05,
"loss": 0.2706,
"step": 103
},
{
"epoch": 52.0,
"learning_rate": 4.771929824561404e-05,
"loss": 0.2642,
"step": 104
},
{
"epoch": 52.5,
"learning_rate": 4.736842105263158e-05,
"loss": 0.2654,
"step": 105
},
{
"epoch": 53.0,
"learning_rate": 4.701754385964913e-05,
"loss": 0.2615,
"step": 106
},
{
"epoch": 53.5,
"learning_rate": 4.666666666666667e-05,
"loss": 0.2564,
"step": 107
},
{
"epoch": 54.0,
"learning_rate": 4.6315789473684214e-05,
"loss": 0.2746,
"step": 108
},
{
"epoch": 54.5,
"learning_rate": 4.5964912280701756e-05,
"loss": 0.2581,
"step": 109
},
{
"epoch": 55.0,
"learning_rate": 4.5614035087719304e-05,
"loss": 0.2593,
"step": 110
},
{
"epoch": 55.0,
"eval_accuracy_dropoff": 0.33773251680084404,
"eval_accuracy_undropoff": 0.9820135187173438,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.23874827725982975,
"eval_iou_undropoff": 0.954587081777339,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.30832570791244507,
"eval_mean_accuracy": 0.6598730177590939,
"eval_mean_iou": 0.5966676795185843,
"eval_overall_accuracy": 0.9552248001098633,
"eval_runtime": 2.8284,
"eval_samples_per_second": 7.071,
"eval_steps_per_second": 0.707,
"step": 110
},
{
"epoch": 55.5,
"learning_rate": 4.5263157894736846e-05,
"loss": 0.2512,
"step": 111
},
{
"epoch": 56.0,
"learning_rate": 4.491228070175439e-05,
"loss": 0.2511,
"step": 112
},
{
"epoch": 56.5,
"learning_rate": 4.456140350877193e-05,
"loss": 0.2618,
"step": 113
},
{
"epoch": 57.0,
"learning_rate": 4.421052631578948e-05,
"loss": 0.2434,
"step": 114
},
{
"epoch": 57.5,
"learning_rate": 4.385964912280702e-05,
"loss": 0.2517,
"step": 115
},
{
"epoch": 58.0,
"learning_rate": 4.350877192982457e-05,
"loss": 0.2416,
"step": 116
},
{
"epoch": 58.5,
"learning_rate": 4.315789473684211e-05,
"loss": 0.244,
"step": 117
},
{
"epoch": 59.0,
"learning_rate": 4.2807017543859654e-05,
"loss": 0.2488,
"step": 118
},
{
"epoch": 59.5,
"learning_rate": 4.2456140350877196e-05,
"loss": 0.2401,
"step": 119
},
{
"epoch": 60.0,
"learning_rate": 4.210526315789474e-05,
"loss": 0.2293,
"step": 120
},
{
"epoch": 60.0,
"eval_accuracy_dropoff": 0.28798825661139016,
"eval_accuracy_undropoff": 0.9897756864087437,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.23306146541387157,
"eval_iou_undropoff": 0.9601183374838324,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.2761629819869995,
"eval_mean_accuracy": 0.6388819715100669,
"eval_mean_iou": 0.5965899014488519,
"eval_overall_accuracy": 0.9605958938598633,
"eval_runtime": 2.9048,
"eval_samples_per_second": 6.885,
"eval_steps_per_second": 0.689,
"step": 120
}
],
"max_steps": 240,
"num_train_epochs": 120,
"total_flos": 6.107874830843904e+16,
"trial_name": null,
"trial_params": null
}