sam1120's picture
Training in progress, step 80
2b69eb2
raw
history blame
14.8 kB
{
"best_metric": 0.1999509036540985,
"best_model_checkpoint": "/robodata/smodak/Projects/nspl/scripts/terrainseg/training/models/dropoff-utcustom-train-SF-RGB-b5_7/checkpoint-80",
"epoch": 40.0,
"global_step": 80,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.5,
"learning_rate": 4.166666666666667e-06,
"loss": 1.017,
"step": 1
},
{
"epoch": 1.0,
"learning_rate": 8.333333333333334e-06,
"loss": 1.025,
"step": 2
},
{
"epoch": 1.5,
"learning_rate": 1.25e-05,
"loss": 1.0124,
"step": 3
},
{
"epoch": 2.0,
"learning_rate": 1.6666666666666667e-05,
"loss": 0.9968,
"step": 4
},
{
"epoch": 2.5,
"learning_rate": 2.0833333333333336e-05,
"loss": 0.9685,
"step": 5
},
{
"epoch": 3.0,
"learning_rate": 2.5e-05,
"loss": 0.9552,
"step": 6
},
{
"epoch": 3.5,
"learning_rate": 2.916666666666667e-05,
"loss": 0.9261,
"step": 7
},
{
"epoch": 4.0,
"learning_rate": 3.3333333333333335e-05,
"loss": 0.8902,
"step": 8
},
{
"epoch": 4.5,
"learning_rate": 3.7500000000000003e-05,
"loss": 0.8624,
"step": 9
},
{
"epoch": 5.0,
"learning_rate": 4.166666666666667e-05,
"loss": 0.8255,
"step": 10
},
{
"epoch": 5.0,
"eval_accuracy_dropoff": 0.61789949310764,
"eval_accuracy_undropoff": 0.9532616169325269,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.30068083396580203,
"eval_iou_undropoff": 0.9377191953832399,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.7949008941650391,
"eval_mean_accuracy": 0.7855805550200834,
"eval_mean_iou": 0.41280000978301395,
"eval_overall_accuracy": 0.9393175125122071,
"eval_runtime": 2.0065,
"eval_samples_per_second": 9.968,
"eval_steps_per_second": 0.997,
"step": 10
},
{
"epoch": 5.5,
"learning_rate": 4.5833333333333334e-05,
"loss": 0.7939,
"step": 11
},
{
"epoch": 6.0,
"learning_rate": 5e-05,
"loss": 0.7364,
"step": 12
},
{
"epoch": 6.5,
"learning_rate": 4.9780701754385966e-05,
"loss": 0.6993,
"step": 13
},
{
"epoch": 7.0,
"learning_rate": 4.956140350877193e-05,
"loss": 0.6614,
"step": 14
},
{
"epoch": 7.5,
"learning_rate": 4.9342105263157894e-05,
"loss": 0.5986,
"step": 15
},
{
"epoch": 8.0,
"learning_rate": 4.912280701754386e-05,
"loss": 0.5876,
"step": 16
},
{
"epoch": 8.5,
"learning_rate": 4.890350877192983e-05,
"loss": 0.5312,
"step": 17
},
{
"epoch": 9.0,
"learning_rate": 4.868421052631579e-05,
"loss": 0.5088,
"step": 18
},
{
"epoch": 9.5,
"learning_rate": 4.8464912280701755e-05,
"loss": 0.462,
"step": 19
},
{
"epoch": 10.0,
"learning_rate": 4.824561403508772e-05,
"loss": 0.4434,
"step": 20
},
{
"epoch": 10.0,
"eval_accuracy_dropoff": 0.41871602559691734,
"eval_accuracy_undropoff": 0.9944325094007127,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.3713824207926633,
"eval_iou_undropoff": 0.9699718650625573,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.42470020055770874,
"eval_mean_accuracy": 0.706574267498815,
"eval_mean_iou": 0.4471180952850735,
"eval_overall_accuracy": 0.9704946517944336,
"eval_runtime": 2.0254,
"eval_samples_per_second": 9.875,
"eval_steps_per_second": 0.987,
"step": 20
},
{
"epoch": 10.5,
"learning_rate": 4.802631578947368e-05,
"loss": 0.3698,
"step": 21
},
{
"epoch": 11.0,
"learning_rate": 4.780701754385965e-05,
"loss": 0.4431,
"step": 22
},
{
"epoch": 11.5,
"learning_rate": 4.758771929824562e-05,
"loss": 0.3667,
"step": 23
},
{
"epoch": 12.0,
"learning_rate": 4.736842105263158e-05,
"loss": 0.3123,
"step": 24
},
{
"epoch": 12.5,
"learning_rate": 4.714912280701755e-05,
"loss": 0.3021,
"step": 25
},
{
"epoch": 13.0,
"learning_rate": 4.6929824561403515e-05,
"loss": 0.2765,
"step": 26
},
{
"epoch": 13.5,
"learning_rate": 4.671052631578948e-05,
"loss": 0.2547,
"step": 27
},
{
"epoch": 14.0,
"learning_rate": 4.649122807017544e-05,
"loss": 0.2615,
"step": 28
},
{
"epoch": 14.5,
"learning_rate": 4.6271929824561406e-05,
"loss": 0.2296,
"step": 29
},
{
"epoch": 15.0,
"learning_rate": 4.605263157894737e-05,
"loss": 0.2107,
"step": 30
},
{
"epoch": 15.0,
"eval_accuracy_dropoff": 0.40458267391453934,
"eval_accuracy_undropoff": 0.9961075726111145,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.3712713740875759,
"eval_iou_undropoff": 0.9710249883455491,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.2725875973701477,
"eval_mean_accuracy": 0.7003451232628269,
"eval_mean_iou": 0.6711481812165625,
"eval_overall_accuracy": 0.9715124130249023,
"eval_runtime": 2.0394,
"eval_samples_per_second": 9.807,
"eval_steps_per_second": 0.981,
"step": 30
},
{
"epoch": 15.5,
"learning_rate": 4.5833333333333334e-05,
"loss": 0.205,
"step": 31
},
{
"epoch": 16.0,
"learning_rate": 4.56140350877193e-05,
"loss": 0.1816,
"step": 32
},
{
"epoch": 16.5,
"learning_rate": 4.539473684210527e-05,
"loss": 0.1842,
"step": 33
},
{
"epoch": 17.0,
"learning_rate": 4.517543859649123e-05,
"loss": 0.149,
"step": 34
},
{
"epoch": 17.5,
"learning_rate": 4.4956140350877196e-05,
"loss": 0.1361,
"step": 35
},
{
"epoch": 18.0,
"learning_rate": 4.473684210526316e-05,
"loss": 0.1646,
"step": 36
},
{
"epoch": 18.5,
"learning_rate": 4.451754385964912e-05,
"loss": 0.1288,
"step": 37
},
{
"epoch": 19.0,
"learning_rate": 4.429824561403509e-05,
"loss": 0.1178,
"step": 38
},
{
"epoch": 19.5,
"learning_rate": 4.407894736842105e-05,
"loss": 0.1103,
"step": 39
},
{
"epoch": 20.0,
"learning_rate": 4.3859649122807014e-05,
"loss": 0.1678,
"step": 40
},
{
"epoch": 20.0,
"eval_accuracy_dropoff": 0.47821739030711713,
"eval_accuracy_undropoff": 0.9904238604465575,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.39174564284479585,
"eval_iou_undropoff": 0.9685003651750826,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.23879647254943848,
"eval_mean_accuracy": 0.7343206253768373,
"eval_mean_iou": 0.6801230040099393,
"eval_overall_accuracy": 0.9691267013549805,
"eval_runtime": 2.833,
"eval_samples_per_second": 7.06,
"eval_steps_per_second": 0.706,
"step": 40
},
{
"epoch": 20.5,
"learning_rate": 4.3640350877192985e-05,
"loss": 0.0934,
"step": 41
},
{
"epoch": 21.0,
"learning_rate": 4.342105263157895e-05,
"loss": 0.1108,
"step": 42
},
{
"epoch": 21.5,
"learning_rate": 4.320175438596491e-05,
"loss": 0.0932,
"step": 43
},
{
"epoch": 22.0,
"learning_rate": 4.298245614035088e-05,
"loss": 0.097,
"step": 44
},
{
"epoch": 22.5,
"learning_rate": 4.2763157894736847e-05,
"loss": 0.0813,
"step": 45
},
{
"epoch": 23.0,
"learning_rate": 4.254385964912281e-05,
"loss": 0.1387,
"step": 46
},
{
"epoch": 23.5,
"learning_rate": 4.2324561403508774e-05,
"loss": 0.0835,
"step": 47
},
{
"epoch": 24.0,
"learning_rate": 4.210526315789474e-05,
"loss": 0.0745,
"step": 48
},
{
"epoch": 24.5,
"learning_rate": 4.188596491228071e-05,
"loss": 0.0736,
"step": 49
},
{
"epoch": 25.0,
"learning_rate": 4.166666666666667e-05,
"loss": 0.0972,
"step": 50
},
{
"epoch": 25.0,
"eval_accuracy_dropoff": 0.42408770843367966,
"eval_accuracy_undropoff": 0.9951975020323848,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.3818202843962053,
"eval_iou_undropoff": 0.9709387427964766,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.18494448065757751,
"eval_mean_accuracy": 0.7096426052330322,
"eval_mean_iou": 0.676379513596341,
"eval_overall_accuracy": 0.9714511871337891,
"eval_runtime": 2.046,
"eval_samples_per_second": 9.775,
"eval_steps_per_second": 0.978,
"step": 50
},
{
"epoch": 25.5,
"learning_rate": 4.1447368421052636e-05,
"loss": 0.0686,
"step": 51
},
{
"epoch": 26.0,
"learning_rate": 4.12280701754386e-05,
"loss": 0.0779,
"step": 52
},
{
"epoch": 26.5,
"learning_rate": 4.100877192982456e-05,
"loss": 0.0723,
"step": 53
},
{
"epoch": 27.0,
"learning_rate": 4.078947368421053e-05,
"loss": 0.0685,
"step": 54
},
{
"epoch": 27.5,
"learning_rate": 4.057017543859649e-05,
"loss": 0.0662,
"step": 55
},
{
"epoch": 28.0,
"learning_rate": 4.0350877192982455e-05,
"loss": 0.0641,
"step": 56
},
{
"epoch": 28.5,
"learning_rate": 4.0131578947368425e-05,
"loss": 0.0567,
"step": 57
},
{
"epoch": 29.0,
"learning_rate": 3.991228070175439e-05,
"loss": 0.0769,
"step": 58
},
{
"epoch": 29.5,
"learning_rate": 3.969298245614035e-05,
"loss": 0.0618,
"step": 59
},
{
"epoch": 30.0,
"learning_rate": 3.9473684210526316e-05,
"loss": 0.0604,
"step": 60
},
{
"epoch": 30.0,
"eval_accuracy_dropoff": 0.5239248606619418,
"eval_accuracy_undropoff": 0.9897283221407057,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.42364186677942717,
"eval_iou_undropoff": 0.9697005089236638,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.20193985104560852,
"eval_mean_accuracy": 0.7568265914013237,
"eval_mean_iou": 0.464447458567697,
"eval_overall_accuracy": 0.9703605651855469,
"eval_runtime": 2.0527,
"eval_samples_per_second": 9.743,
"eval_steps_per_second": 0.974,
"step": 60
},
{
"epoch": 30.5,
"learning_rate": 3.925438596491228e-05,
"loss": 0.0556,
"step": 61
},
{
"epoch": 31.0,
"learning_rate": 3.9035087719298244e-05,
"loss": 0.0561,
"step": 62
},
{
"epoch": 31.5,
"learning_rate": 3.8815789473684214e-05,
"loss": 0.055,
"step": 63
},
{
"epoch": 32.0,
"learning_rate": 3.859649122807018e-05,
"loss": 0.0524,
"step": 64
},
{
"epoch": 32.5,
"learning_rate": 3.837719298245614e-05,
"loss": 0.051,
"step": 65
},
{
"epoch": 33.0,
"learning_rate": 3.815789473684211e-05,
"loss": 0.0542,
"step": 66
},
{
"epoch": 33.5,
"learning_rate": 3.7938596491228076e-05,
"loss": 0.0516,
"step": 67
},
{
"epoch": 34.0,
"learning_rate": 3.771929824561404e-05,
"loss": 0.0466,
"step": 68
},
{
"epoch": 34.5,
"learning_rate": 3.7500000000000003e-05,
"loss": 0.0502,
"step": 69
},
{
"epoch": 35.0,
"learning_rate": 3.728070175438597e-05,
"loss": 0.0497,
"step": 70
},
{
"epoch": 35.0,
"eval_accuracy_dropoff": 0.4775384756531113,
"eval_accuracy_undropoff": 0.9913645784928411,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.39826387693343585,
"eval_iou_undropoff": 0.9693923400870129,
"eval_iou_unlabeled": NaN,
"eval_loss": 0.1793430596590042,
"eval_mean_accuracy": 0.7344515270729762,
"eval_mean_iou": 0.6838281085102244,
"eval_overall_accuracy": 0.9700000762939454,
"eval_runtime": 6.1591,
"eval_samples_per_second": 3.247,
"eval_steps_per_second": 0.325,
"step": 70
},
{
"epoch": 35.5,
"learning_rate": 3.706140350877193e-05,
"loss": 0.0442,
"step": 71
},
{
"epoch": 36.0,
"learning_rate": 3.6842105263157895e-05,
"loss": 0.0477,
"step": 72
},
{
"epoch": 36.5,
"learning_rate": 3.662280701754386e-05,
"loss": 0.0418,
"step": 73
},
{
"epoch": 37.0,
"learning_rate": 3.640350877192983e-05,
"loss": 0.0829,
"step": 74
},
{
"epoch": 37.5,
"learning_rate": 3.618421052631579e-05,
"loss": 0.0458,
"step": 75
},
{
"epoch": 38.0,
"learning_rate": 3.5964912280701756e-05,
"loss": 0.0465,
"step": 76
},
{
"epoch": 38.5,
"learning_rate": 3.574561403508772e-05,
"loss": 0.0445,
"step": 77
},
{
"epoch": 39.0,
"learning_rate": 3.5526315789473684e-05,
"loss": 0.0509,
"step": 78
},
{
"epoch": 39.5,
"learning_rate": 3.530701754385965e-05,
"loss": 0.0429,
"step": 79
},
{
"epoch": 40.0,
"learning_rate": 3.508771929824561e-05,
"loss": 0.0492,
"step": 80
},
{
"epoch": 40.0,
"eval_accuracy_dropoff": 0.5238606390054817,
"eval_accuracy_undropoff": 0.9895599600786884,
"eval_accuracy_unlabeled": NaN,
"eval_iou_dropoff": 0.42225237750136807,
"eval_iou_undropoff": 0.969532907193877,
"eval_iou_unlabeled": 0.0,
"eval_loss": 0.1999509036540985,
"eval_mean_accuracy": 0.7567102995420851,
"eval_mean_iou": 0.46392842823174835,
"eval_overall_accuracy": 0.970196533203125,
"eval_runtime": 2.1001,
"eval_samples_per_second": 9.523,
"eval_steps_per_second": 0.952,
"step": 80
}
],
"max_steps": 240,
"num_train_epochs": 120,
"total_flos": 4.630399399703347e+17,
"trial_name": null,
"trial_params": null
}