sam1120's picture
Training in progress, step 120
da2c14d
raw
history blame
22.1 kB
{
"best_metric": 0.4159059226512909,
"best_model_checkpoint": "/robodata/smodak/Projects/nspl/scripts/terrainseg/training/models/dropoff-utcustom-train-SF-RGB-b0_5/checkpoint-120",
"epoch": 40.0,
"global_step": 120,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.33,
"learning_rate": 2.777777777777778e-06,
"loss": 1.249,
"step": 1
},
{
"epoch": 0.67,
"learning_rate": 5.555555555555556e-06,
"loss": 1.2488,
"step": 2
},
{
"epoch": 1.0,
"learning_rate": 8.333333333333334e-06,
"loss": 1.2553,
"step": 3
},
{
"epoch": 1.33,
"learning_rate": 1.1111111111111112e-05,
"loss": 1.2475,
"step": 4
},
{
"epoch": 1.67,
"learning_rate": 1.388888888888889e-05,
"loss": 1.2427,
"step": 5
},
{
"epoch": 2.0,
"learning_rate": 1.6666666666666667e-05,
"loss": 1.2497,
"step": 6
},
{
"epoch": 2.33,
"learning_rate": 1.9444444444444445e-05,
"loss": 1.2339,
"step": 7
},
{
"epoch": 2.67,
"learning_rate": 2.2222222222222223e-05,
"loss": 1.2302,
"step": 8
},
{
"epoch": 3.0,
"learning_rate": 2.5e-05,
"loss": 1.2196,
"step": 9
},
{
"epoch": 3.33,
"learning_rate": 2.777777777777778e-05,
"loss": 1.2123,
"step": 10
},
{
"epoch": 3.33,
"eval_accuracy_dropoff": 0.1908240861993705,
"eval_accuracy_undropoff": 0.18874432817853504,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.04942539850194509,
"eval_iou_undropoff": 0.188595832344016,
"eval_iou_unlabeled": 0.0,
"eval_loss": 1.1206141710281372,
"eval_mean_accuracy": 0.18978420718895278,
"eval_mean_iou": 0.07934041028198703,
"eval_overall_accuracy": 0.18883641560872397,
"eval_runtime": 1.7856,
"eval_samples_per_second": 8.401,
"eval_steps_per_second": 0.56,
"step": 10
},
{
"epoch": 3.67,
"learning_rate": 3.055555555555556e-05,
"loss": 1.2065,
"step": 11
},
{
"epoch": 4.0,
"learning_rate": 3.3333333333333335e-05,
"loss": 1.1867,
"step": 12
},
{
"epoch": 4.33,
"learning_rate": 3.611111111111111e-05,
"loss": 1.1956,
"step": 13
},
{
"epoch": 4.67,
"learning_rate": 3.888888888888889e-05,
"loss": 1.1767,
"step": 14
},
{
"epoch": 5.0,
"learning_rate": 4.166666666666667e-05,
"loss": 1.1659,
"step": 15
},
{
"epoch": 5.33,
"learning_rate": 4.4444444444444447e-05,
"loss": 1.1494,
"step": 16
},
{
"epoch": 5.67,
"learning_rate": 4.722222222222222e-05,
"loss": 1.1325,
"step": 17
},
{
"epoch": 6.0,
"learning_rate": 5e-05,
"loss": 1.1703,
"step": 18
},
{
"epoch": 6.33,
"learning_rate": 4.985380116959065e-05,
"loss": 1.1172,
"step": 19
},
{
"epoch": 6.67,
"learning_rate": 4.970760233918128e-05,
"loss": 1.0927,
"step": 20
},
{
"epoch": 6.67,
"eval_accuracy_dropoff": 0.6449789785650286,
"eval_accuracy_undropoff": 0.5299809582198437,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.12901550653889501,
"eval_iou_undropoff": 0.529796420189495,
"eval_iou_unlabeled": 0.0,
"eval_loss": 1.0984647274017334,
"eval_mean_accuracy": 0.5874799683924361,
"eval_mean_iou": 0.21960397557613,
"eval_overall_accuracy": 0.5350728352864583,
"eval_runtime": 1.7244,
"eval_samples_per_second": 8.699,
"eval_steps_per_second": 0.58,
"step": 20
},
{
"epoch": 7.0,
"learning_rate": 4.956140350877193e-05,
"loss": 1.1371,
"step": 21
},
{
"epoch": 7.33,
"learning_rate": 4.941520467836258e-05,
"loss": 1.064,
"step": 22
},
{
"epoch": 7.67,
"learning_rate": 4.926900584795322e-05,
"loss": 1.0567,
"step": 23
},
{
"epoch": 8.0,
"learning_rate": 4.912280701754386e-05,
"loss": 1.0585,
"step": 24
},
{
"epoch": 8.33,
"learning_rate": 4.8976608187134504e-05,
"loss": 1.0397,
"step": 25
},
{
"epoch": 8.67,
"learning_rate": 4.883040935672515e-05,
"loss": 1.0011,
"step": 26
},
{
"epoch": 9.0,
"learning_rate": 4.868421052631579e-05,
"loss": 0.9476,
"step": 27
},
{
"epoch": 9.33,
"learning_rate": 4.853801169590643e-05,
"loss": 0.9694,
"step": 28
},
{
"epoch": 9.67,
"learning_rate": 4.839181286549708e-05,
"loss": 0.9674,
"step": 29
},
{
"epoch": 10.0,
"learning_rate": 4.824561403508772e-05,
"loss": 1.0578,
"step": 30
},
{
"epoch": 10.0,
"eval_accuracy_dropoff": 0.6399533622808832,
"eval_accuracy_undropoff": 0.8725462021281238,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.23667468886542886,
"eval_iou_undropoff": 0.8620580507674747,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.9785888195037842,
"eval_mean_accuracy": 0.7562497822045036,
"eval_mean_iou": 0.36624424654430116,
"eval_overall_accuracy": 0.8622474670410156,
"eval_runtime": 1.8276,
"eval_samples_per_second": 8.208,
"eval_steps_per_second": 0.547,
"step": 30
},
{
"epoch": 10.33,
"learning_rate": 4.8099415204678366e-05,
"loss": 0.9272,
"step": 31
},
{
"epoch": 10.67,
"learning_rate": 4.7953216374269006e-05,
"loss": 0.9222,
"step": 32
},
{
"epoch": 11.0,
"learning_rate": 4.780701754385965e-05,
"loss": 0.8411,
"step": 33
},
{
"epoch": 11.33,
"learning_rate": 4.7660818713450294e-05,
"loss": 0.889,
"step": 34
},
{
"epoch": 11.67,
"learning_rate": 4.751461988304094e-05,
"loss": 0.8438,
"step": 35
},
{
"epoch": 12.0,
"learning_rate": 4.736842105263158e-05,
"loss": 0.86,
"step": 36
},
{
"epoch": 12.33,
"learning_rate": 4.722222222222222e-05,
"loss": 0.8268,
"step": 37
},
{
"epoch": 12.67,
"learning_rate": 4.707602339181287e-05,
"loss": 0.8049,
"step": 38
},
{
"epoch": 13.0,
"learning_rate": 4.6929824561403515e-05,
"loss": 1.0057,
"step": 39
},
{
"epoch": 13.33,
"learning_rate": 4.678362573099415e-05,
"loss": 0.788,
"step": 40
},
{
"epoch": 13.33,
"eval_accuracy_dropoff": 0.536460128196292,
"eval_accuracy_undropoff": 0.964590165330336,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.3397820203137278,
"eval_iou_undropoff": 0.9467918336729614,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.7939841747283936,
"eval_mean_accuracy": 0.750525146763314,
"eval_mean_iou": 0.42885795132889637,
"eval_overall_accuracy": 0.9456334431966146,
"eval_runtime": 1.891,
"eval_samples_per_second": 7.932,
"eval_steps_per_second": 0.529,
"step": 40
},
{
"epoch": 13.67,
"learning_rate": 4.6637426900584796e-05,
"loss": 0.7872,
"step": 41
},
{
"epoch": 14.0,
"learning_rate": 4.649122807017544e-05,
"loss": 0.8197,
"step": 42
},
{
"epoch": 14.33,
"learning_rate": 4.634502923976608e-05,
"loss": 0.7351,
"step": 43
},
{
"epoch": 14.67,
"learning_rate": 4.619883040935672e-05,
"loss": 0.7447,
"step": 44
},
{
"epoch": 15.0,
"learning_rate": 4.605263157894737e-05,
"loss": 0.6917,
"step": 45
},
{
"epoch": 15.33,
"learning_rate": 4.590643274853802e-05,
"loss": 0.7277,
"step": 46
},
{
"epoch": 15.67,
"learning_rate": 4.576023391812866e-05,
"loss": 0.6602,
"step": 47
},
{
"epoch": 16.0,
"learning_rate": 4.56140350877193e-05,
"loss": 0.6809,
"step": 48
},
{
"epoch": 16.33,
"learning_rate": 4.5467836257309945e-05,
"loss": 0.6865,
"step": 49
},
{
"epoch": 16.67,
"learning_rate": 4.5321637426900585e-05,
"loss": 0.6353,
"step": 50
},
{
"epoch": 16.67,
"eval_accuracy_dropoff": 0.38302088358949615,
"eval_accuracy_undropoff": 0.9849956839341233,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.2966305629072793,
"eval_iou_undropoff": 0.9581110717398175,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.620606541633606,
"eval_mean_accuracy": 0.6840082837618098,
"eval_mean_iou": 0.41824721154903227,
"eval_overall_accuracy": 0.9583414713541667,
"eval_runtime": 1.7715,
"eval_samples_per_second": 8.467,
"eval_steps_per_second": 0.564,
"step": 50
},
{
"epoch": 17.0,
"learning_rate": 4.517543859649123e-05,
"loss": 0.6601,
"step": 51
},
{
"epoch": 17.33,
"learning_rate": 4.502923976608187e-05,
"loss": 0.6353,
"step": 52
},
{
"epoch": 17.67,
"learning_rate": 4.488304093567251e-05,
"loss": 0.6038,
"step": 53
},
{
"epoch": 18.0,
"learning_rate": 4.473684210526316e-05,
"loss": 0.5588,
"step": 54
},
{
"epoch": 18.33,
"learning_rate": 4.4590643274853806e-05,
"loss": 0.5965,
"step": 55
},
{
"epoch": 18.67,
"learning_rate": 4.4444444444444447e-05,
"loss": 0.5992,
"step": 56
},
{
"epoch": 19.0,
"learning_rate": 4.429824561403509e-05,
"loss": 0.7979,
"step": 57
},
{
"epoch": 19.33,
"learning_rate": 4.4152046783625734e-05,
"loss": 0.5471,
"step": 58
},
{
"epoch": 19.67,
"learning_rate": 4.400584795321638e-05,
"loss": 0.5725,
"step": 59
},
{
"epoch": 20.0,
"learning_rate": 4.3859649122807014e-05,
"loss": 0.6944,
"step": 60
},
{
"epoch": 20.0,
"eval_accuracy_dropoff": 0.36309072529694214,
"eval_accuracy_undropoff": 0.9900834262005954,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.30135334188209384,
"eval_iou_undropoff": 0.9619544548614083,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.52126544713974,
"eval_mean_accuracy": 0.6765870757487688,
"eval_mean_iou": 0.4211025989145007,
"eval_overall_accuracy": 0.9623214721679687,
"eval_runtime": 1.857,
"eval_samples_per_second": 8.078,
"eval_steps_per_second": 0.539,
"step": 60
},
{
"epoch": 20.33,
"learning_rate": 4.371345029239766e-05,
"loss": 0.5386,
"step": 61
},
{
"epoch": 20.67,
"learning_rate": 4.356725146198831e-05,
"loss": 0.5411,
"step": 62
},
{
"epoch": 21.0,
"learning_rate": 4.342105263157895e-05,
"loss": 0.8263,
"step": 63
},
{
"epoch": 21.33,
"learning_rate": 4.327485380116959e-05,
"loss": 0.556,
"step": 64
},
{
"epoch": 21.67,
"learning_rate": 4.3128654970760236e-05,
"loss": 0.5597,
"step": 65
},
{
"epoch": 22.0,
"learning_rate": 4.298245614035088e-05,
"loss": 0.5996,
"step": 66
},
{
"epoch": 22.33,
"learning_rate": 4.283625730994152e-05,
"loss": 0.5711,
"step": 67
},
{
"epoch": 22.67,
"learning_rate": 4.269005847953216e-05,
"loss": 0.5204,
"step": 68
},
{
"epoch": 23.0,
"learning_rate": 4.254385964912281e-05,
"loss": 0.8124,
"step": 69
},
{
"epoch": 23.33,
"learning_rate": 4.239766081871345e-05,
"loss": 0.5046,
"step": 70
},
{
"epoch": 23.33,
"eval_accuracy_dropoff": 0.36825418705630986,
"eval_accuracy_undropoff": 0.9909756437643757,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.30895554752439464,
"eval_iou_undropoff": 0.9628048952316295,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.4764857590198517,
"eval_mean_accuracy": 0.6796149154103428,
"eval_mean_iou": 0.42392014758534136,
"eval_overall_accuracy": 0.9634028116861979,
"eval_runtime": 1.754,
"eval_samples_per_second": 8.552,
"eval_steps_per_second": 0.57,
"step": 70
},
{
"epoch": 23.67,
"learning_rate": 4.22514619883041e-05,
"loss": 0.5107,
"step": 71
},
{
"epoch": 24.0,
"learning_rate": 4.210526315789474e-05,
"loss": 0.5692,
"step": 72
},
{
"epoch": 24.33,
"learning_rate": 4.195906432748538e-05,
"loss": 0.4992,
"step": 73
},
{
"epoch": 24.67,
"learning_rate": 4.1812865497076025e-05,
"loss": 0.514,
"step": 74
},
{
"epoch": 25.0,
"learning_rate": 4.166666666666667e-05,
"loss": 0.7924,
"step": 75
},
{
"epoch": 25.33,
"learning_rate": 4.152046783625731e-05,
"loss": 0.4816,
"step": 76
},
{
"epoch": 25.67,
"learning_rate": 4.137426900584795e-05,
"loss": 0.4831,
"step": 77
},
{
"epoch": 26.0,
"learning_rate": 4.12280701754386e-05,
"loss": 0.8194,
"step": 78
},
{
"epoch": 26.33,
"learning_rate": 4.1081871345029247e-05,
"loss": 0.5093,
"step": 79
},
{
"epoch": 26.67,
"learning_rate": 4.093567251461988e-05,
"loss": 0.4684,
"step": 80
},
{
"epoch": 26.67,
"eval_accuracy_dropoff": 0.27794242653984885,
"eval_accuracy_undropoff": 0.9914216194986126,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.23522092062411898,
"eval_iou_undropoff": 0.9593306905362942,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.4643230140209198,
"eval_mean_accuracy": 0.6346820230192307,
"eval_mean_iou": 0.39818387038680436,
"eval_overall_accuracy": 0.9598302205403646,
"eval_runtime": 1.7927,
"eval_samples_per_second": 8.367,
"eval_steps_per_second": 0.558,
"step": 80
},
{
"epoch": 27.0,
"learning_rate": 4.078947368421053e-05,
"loss": 0.5992,
"step": 81
},
{
"epoch": 27.33,
"learning_rate": 4.0643274853801174e-05,
"loss": 0.505,
"step": 82
},
{
"epoch": 27.67,
"learning_rate": 4.0497076023391814e-05,
"loss": 0.4917,
"step": 83
},
{
"epoch": 28.0,
"learning_rate": 4.0350877192982455e-05,
"loss": 0.4608,
"step": 84
},
{
"epoch": 28.33,
"learning_rate": 4.02046783625731e-05,
"loss": 0.4843,
"step": 85
},
{
"epoch": 28.67,
"learning_rate": 4.005847953216375e-05,
"loss": 0.4416,
"step": 86
},
{
"epoch": 29.0,
"learning_rate": 3.991228070175439e-05,
"loss": 0.7439,
"step": 87
},
{
"epoch": 29.33,
"learning_rate": 3.976608187134503e-05,
"loss": 0.4523,
"step": 88
},
{
"epoch": 29.67,
"learning_rate": 3.9619883040935676e-05,
"loss": 0.4525,
"step": 89
},
{
"epoch": 30.0,
"learning_rate": 3.9473684210526316e-05,
"loss": 0.4401,
"step": 90
},
{
"epoch": 30.0,
"eval_accuracy_dropoff": 0.3077400234337308,
"eval_accuracy_undropoff": 0.9935777897698063,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.2703275413210631,
"eval_iou_undropoff": 0.9627020713766379,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.44831418991088867,
"eval_mean_accuracy": 0.6506589066017685,
"eval_mean_iou": 0.41100987089923363,
"eval_overall_accuracy": 0.9632102966308593,
"eval_runtime": 1.8561,
"eval_samples_per_second": 8.082,
"eval_steps_per_second": 0.539,
"step": 90
},
{
"epoch": 30.33,
"learning_rate": 3.932748538011696e-05,
"loss": 0.4473,
"step": 91
},
{
"epoch": 30.67,
"learning_rate": 3.9181286549707604e-05,
"loss": 0.4753,
"step": 92
},
{
"epoch": 31.0,
"learning_rate": 3.9035087719298244e-05,
"loss": 0.7232,
"step": 93
},
{
"epoch": 31.33,
"learning_rate": 3.888888888888889e-05,
"loss": 0.4391,
"step": 94
},
{
"epoch": 31.67,
"learning_rate": 3.874269005847954e-05,
"loss": 0.4428,
"step": 95
},
{
"epoch": 32.0,
"learning_rate": 3.859649122807018e-05,
"loss": 0.457,
"step": 96
},
{
"epoch": 32.33,
"learning_rate": 3.845029239766082e-05,
"loss": 0.4326,
"step": 97
},
{
"epoch": 32.67,
"learning_rate": 3.8304093567251465e-05,
"loss": 0.4359,
"step": 98
},
{
"epoch": 33.0,
"learning_rate": 3.815789473684211e-05,
"loss": 0.7037,
"step": 99
},
{
"epoch": 33.33,
"learning_rate": 3.8011695906432746e-05,
"loss": 0.4268,
"step": 100
},
{
"epoch": 33.33,
"eval_accuracy_dropoff": 0.4108484388999931,
"eval_accuracy_undropoff": 0.9894514498468887,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.33465263157894737,
"eval_iou_undropoff": 0.9631619518896926,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.4365720748901367,
"eval_mean_accuracy": 0.7001499443734409,
"eval_mean_iou": 0.6489072917343199,
"eval_overall_accuracy": 0.9638320922851562,
"eval_runtime": 1.7768,
"eval_samples_per_second": 8.442,
"eval_steps_per_second": 0.563,
"step": 100
},
{
"epoch": 33.67,
"learning_rate": 3.786549707602339e-05,
"loss": 0.4203,
"step": 101
},
{
"epoch": 34.0,
"learning_rate": 3.771929824561404e-05,
"loss": 0.6864,
"step": 102
},
{
"epoch": 34.33,
"learning_rate": 3.757309941520468e-05,
"loss": 0.4093,
"step": 103
},
{
"epoch": 34.67,
"learning_rate": 3.742690058479532e-05,
"loss": 0.4149,
"step": 104
},
{
"epoch": 35.0,
"learning_rate": 3.728070175438597e-05,
"loss": 0.3847,
"step": 105
},
{
"epoch": 35.33,
"learning_rate": 3.713450292397661e-05,
"loss": 0.4227,
"step": 106
},
{
"epoch": 35.67,
"learning_rate": 3.6988304093567254e-05,
"loss": 0.3916,
"step": 107
},
{
"epoch": 36.0,
"learning_rate": 3.6842105263157895e-05,
"loss": 0.6716,
"step": 108
},
{
"epoch": 36.33,
"learning_rate": 3.669590643274854e-05,
"loss": 0.4096,
"step": 109
},
{
"epoch": 36.67,
"learning_rate": 3.654970760233918e-05,
"loss": 0.3939,
"step": 110
},
{
"epoch": 36.67,
"eval_accuracy_dropoff": 0.36697911641050385,
"eval_accuracy_undropoff": 0.9927164392616175,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.3171232876712329,
"eval_iou_undropoff": 0.9644338706379278,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.4026946723461151,
"eval_mean_accuracy": 0.6798477778360607,
"eval_mean_iou": 0.4271857194363869,
"eval_overall_accuracy": 0.9650100708007813,
"eval_runtime": 1.777,
"eval_samples_per_second": 8.441,
"eval_steps_per_second": 0.563,
"step": 110
},
{
"epoch": 37.0,
"learning_rate": 3.640350877192983e-05,
"loss": 0.4204,
"step": 111
},
{
"epoch": 37.33,
"learning_rate": 3.625730994152047e-05,
"loss": 0.4005,
"step": 112
},
{
"epoch": 37.67,
"learning_rate": 3.611111111111111e-05,
"loss": 0.4024,
"step": 113
},
{
"epoch": 38.0,
"learning_rate": 3.5964912280701756e-05,
"loss": 0.6594,
"step": 114
},
{
"epoch": 38.33,
"learning_rate": 3.5818713450292403e-05,
"loss": 0.3816,
"step": 115
},
{
"epoch": 38.67,
"learning_rate": 3.5672514619883044e-05,
"loss": 0.4391,
"step": 116
},
{
"epoch": 39.0,
"learning_rate": 3.5526315789473684e-05,
"loss": 0.4499,
"step": 117
},
{
"epoch": 39.33,
"learning_rate": 3.538011695906433e-05,
"loss": 0.3772,
"step": 118
},
{
"epoch": 39.67,
"learning_rate": 3.523391812865498e-05,
"loss": 0.3751,
"step": 119
},
{
"epoch": 40.0,
"learning_rate": 3.508771929824561e-05,
"loss": 0.4472,
"step": 120
},
{
"epoch": 40.0,
"eval_accuracy_dropoff": 0.38870700944241504,
"eval_accuracy_undropoff": 0.9904841657326722,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.3224725900442657,
"eval_iou_undropoff": 0.963205427478915,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.4159059226512909,
"eval_mean_accuracy": 0.6895955875875436,
"eval_mean_iou": 0.6428390087615904,
"eval_overall_accuracy": 0.9638387044270833,
"eval_runtime": 1.8087,
"eval_samples_per_second": 8.293,
"eval_steps_per_second": 0.553,
"step": 120
}
],
"max_steps": 360,
"num_train_epochs": 120,
"total_flos": 2.38396687515648e+16,
"trial_name": null,
"trial_params": null
}