|
{
|
|
"best_metric": 0.8333333333333334,
|
|
"best_model_checkpoint": "vit-base-patch16-224-RU2-40\\checkpoint-1155",
|
|
"epoch": 39.48051948051948,
|
|
"eval_steps": 500,
|
|
"global_step": 1520,
|
|
"is_hyper_param_search": false,
|
|
"is_local_process_zero": true,
|
|
"is_world_process_zero": true,
|
|
"log_history": [
|
|
{
|
|
"epoch": 0.26,
|
|
"learning_rate": 7.2368421052631575e-06,
|
|
"loss": 1.384,
|
|
"step": 10
|
|
},
|
|
{
|
|
"epoch": 0.52,
|
|
"learning_rate": 1.4473684210526315e-05,
|
|
"loss": 1.3667,
|
|
"step": 20
|
|
},
|
|
{
|
|
"epoch": 0.78,
|
|
"learning_rate": 2.1710526315789474e-05,
|
|
"loss": 1.3226,
|
|
"step": 30
|
|
},
|
|
{
|
|
"epoch": 0.99,
|
|
"eval_accuracy": 0.6,
|
|
"eval_loss": 1.2293227910995483,
|
|
"eval_runtime": 1.0844,
|
|
"eval_samples_per_second": 55.33,
|
|
"eval_steps_per_second": 1.844,
|
|
"step": 38
|
|
},
|
|
{
|
|
"epoch": 1.04,
|
|
"learning_rate": 2.894736842105263e-05,
|
|
"loss": 1.2601,
|
|
"step": 40
|
|
},
|
|
{
|
|
"epoch": 1.3,
|
|
"learning_rate": 3.618421052631579e-05,
|
|
"loss": 1.1706,
|
|
"step": 50
|
|
},
|
|
{
|
|
"epoch": 1.56,
|
|
"learning_rate": 4.342105263157895e-05,
|
|
"loss": 1.0292,
|
|
"step": 60
|
|
},
|
|
{
|
|
"epoch": 1.82,
|
|
"learning_rate": 5.0657894736842104e-05,
|
|
"loss": 0.9048,
|
|
"step": 70
|
|
},
|
|
{
|
|
"epoch": 2.0,
|
|
"eval_accuracy": 0.7,
|
|
"eval_loss": 0.7968690991401672,
|
|
"eval_runtime": 1.0427,
|
|
"eval_samples_per_second": 57.542,
|
|
"eval_steps_per_second": 1.918,
|
|
"step": 77
|
|
},
|
|
{
|
|
"epoch": 2.08,
|
|
"learning_rate": 5.484764542936288e-05,
|
|
"loss": 0.7837,
|
|
"step": 80
|
|
},
|
|
{
|
|
"epoch": 2.34,
|
|
"learning_rate": 5.446675900277008e-05,
|
|
"loss": 0.6251,
|
|
"step": 90
|
|
},
|
|
{
|
|
"epoch": 2.6,
|
|
"learning_rate": 5.408587257617729e-05,
|
|
"loss": 0.4961,
|
|
"step": 100
|
|
},
|
|
{
|
|
"epoch": 2.86,
|
|
"learning_rate": 5.370498614958449e-05,
|
|
"loss": 0.4039,
|
|
"step": 110
|
|
},
|
|
{
|
|
"epoch": 2.99,
|
|
"eval_accuracy": 0.7166666666666667,
|
|
"eval_loss": 0.6800220012664795,
|
|
"eval_runtime": 1.1199,
|
|
"eval_samples_per_second": 53.578,
|
|
"eval_steps_per_second": 1.786,
|
|
"step": 115
|
|
},
|
|
{
|
|
"epoch": 3.12,
|
|
"learning_rate": 5.332409972299169e-05,
|
|
"loss": 0.3074,
|
|
"step": 120
|
|
},
|
|
{
|
|
"epoch": 3.38,
|
|
"learning_rate": 5.294321329639889e-05,
|
|
"loss": 0.2772,
|
|
"step": 130
|
|
},
|
|
{
|
|
"epoch": 3.64,
|
|
"learning_rate": 5.256232686980609e-05,
|
|
"loss": 0.2567,
|
|
"step": 140
|
|
},
|
|
{
|
|
"epoch": 3.9,
|
|
"learning_rate": 5.21814404432133e-05,
|
|
"loss": 0.281,
|
|
"step": 150
|
|
},
|
|
{
|
|
"epoch": 4.0,
|
|
"eval_accuracy": 0.7666666666666667,
|
|
"eval_loss": 0.8892202973365784,
|
|
"eval_runtime": 1.134,
|
|
"eval_samples_per_second": 52.909,
|
|
"eval_steps_per_second": 1.764,
|
|
"step": 154
|
|
},
|
|
{
|
|
"epoch": 4.16,
|
|
"learning_rate": 5.18005540166205e-05,
|
|
"loss": 0.2168,
|
|
"step": 160
|
|
},
|
|
{
|
|
"epoch": 4.42,
|
|
"learning_rate": 5.14196675900277e-05,
|
|
"loss": 0.1975,
|
|
"step": 170
|
|
},
|
|
{
|
|
"epoch": 4.68,
|
|
"learning_rate": 5.1038781163434903e-05,
|
|
"loss": 0.1988,
|
|
"step": 180
|
|
},
|
|
{
|
|
"epoch": 4.94,
|
|
"learning_rate": 5.0657894736842104e-05,
|
|
"loss": 0.1755,
|
|
"step": 190
|
|
},
|
|
{
|
|
"epoch": 4.99,
|
|
"eval_accuracy": 0.7333333333333333,
|
|
"eval_loss": 0.9071826934814453,
|
|
"eval_runtime": 1.08,
|
|
"eval_samples_per_second": 55.555,
|
|
"eval_steps_per_second": 1.852,
|
|
"step": 192
|
|
},
|
|
{
|
|
"epoch": 5.19,
|
|
"learning_rate": 5.027700831024931e-05,
|
|
"loss": 0.1255,
|
|
"step": 200
|
|
},
|
|
{
|
|
"epoch": 5.45,
|
|
"learning_rate": 4.989612188365651e-05,
|
|
"loss": 0.1565,
|
|
"step": 210
|
|
},
|
|
{
|
|
"epoch": 5.71,
|
|
"learning_rate": 4.9515235457063714e-05,
|
|
"loss": 0.1548,
|
|
"step": 220
|
|
},
|
|
{
|
|
"epoch": 5.97,
|
|
"learning_rate": 4.9134349030470915e-05,
|
|
"loss": 0.1035,
|
|
"step": 230
|
|
},
|
|
{
|
|
"epoch": 6.0,
|
|
"eval_accuracy": 0.8166666666666667,
|
|
"eval_loss": 0.8035634756088257,
|
|
"eval_runtime": 1.0846,
|
|
"eval_samples_per_second": 55.318,
|
|
"eval_steps_per_second": 1.844,
|
|
"step": 231
|
|
},
|
|
{
|
|
"epoch": 6.23,
|
|
"learning_rate": 4.8753462603878116e-05,
|
|
"loss": 0.1059,
|
|
"step": 240
|
|
},
|
|
{
|
|
"epoch": 6.49,
|
|
"learning_rate": 4.837257617728532e-05,
|
|
"loss": 0.1615,
|
|
"step": 250
|
|
},
|
|
{
|
|
"epoch": 6.75,
|
|
"learning_rate": 4.7991689750692524e-05,
|
|
"loss": 0.1275,
|
|
"step": 260
|
|
},
|
|
{
|
|
"epoch": 6.99,
|
|
"eval_accuracy": 0.8,
|
|
"eval_loss": 0.8626561760902405,
|
|
"eval_runtime": 1.1485,
|
|
"eval_samples_per_second": 52.243,
|
|
"eval_steps_per_second": 1.741,
|
|
"step": 269
|
|
},
|
|
{
|
|
"epoch": 7.01,
|
|
"learning_rate": 4.7610803324099725e-05,
|
|
"loss": 0.1117,
|
|
"step": 270
|
|
},
|
|
{
|
|
"epoch": 7.27,
|
|
"learning_rate": 4.7229916897506926e-05,
|
|
"loss": 0.0851,
|
|
"step": 280
|
|
},
|
|
{
|
|
"epoch": 7.53,
|
|
"learning_rate": 4.684903047091413e-05,
|
|
"loss": 0.1013,
|
|
"step": 290
|
|
},
|
|
{
|
|
"epoch": 7.79,
|
|
"learning_rate": 4.6468144044321335e-05,
|
|
"loss": 0.107,
|
|
"step": 300
|
|
},
|
|
{
|
|
"epoch": 8.0,
|
|
"eval_accuracy": 0.8,
|
|
"eval_loss": 0.8712760806083679,
|
|
"eval_runtime": 1.0545,
|
|
"eval_samples_per_second": 56.898,
|
|
"eval_steps_per_second": 1.897,
|
|
"step": 308
|
|
},
|
|
{
|
|
"epoch": 8.05,
|
|
"learning_rate": 4.6087257617728535e-05,
|
|
"loss": 0.0832,
|
|
"step": 310
|
|
},
|
|
{
|
|
"epoch": 8.31,
|
|
"learning_rate": 4.570637119113573e-05,
|
|
"loss": 0.087,
|
|
"step": 320
|
|
},
|
|
{
|
|
"epoch": 8.57,
|
|
"learning_rate": 4.532548476454294e-05,
|
|
"loss": 0.0847,
|
|
"step": 330
|
|
},
|
|
{
|
|
"epoch": 8.83,
|
|
"learning_rate": 4.494459833795014e-05,
|
|
"loss": 0.0984,
|
|
"step": 340
|
|
},
|
|
{
|
|
"epoch": 8.99,
|
|
"eval_accuracy": 0.8,
|
|
"eval_loss": 0.965960681438446,
|
|
"eval_runtime": 1.0848,
|
|
"eval_samples_per_second": 55.31,
|
|
"eval_steps_per_second": 1.844,
|
|
"step": 346
|
|
},
|
|
{
|
|
"epoch": 9.09,
|
|
"learning_rate": 4.4563711911357346e-05,
|
|
"loss": 0.082,
|
|
"step": 350
|
|
},
|
|
{
|
|
"epoch": 9.35,
|
|
"learning_rate": 4.418282548476455e-05,
|
|
"loss": 0.0803,
|
|
"step": 360
|
|
},
|
|
{
|
|
"epoch": 9.61,
|
|
"learning_rate": 4.380193905817174e-05,
|
|
"loss": 0.0774,
|
|
"step": 370
|
|
},
|
|
{
|
|
"epoch": 9.87,
|
|
"learning_rate": 4.342105263157895e-05,
|
|
"loss": 0.0823,
|
|
"step": 380
|
|
},
|
|
{
|
|
"epoch": 10.0,
|
|
"eval_accuracy": 0.7833333333333333,
|
|
"eval_loss": 1.070397973060608,
|
|
"eval_runtime": 1.1053,
|
|
"eval_samples_per_second": 54.286,
|
|
"eval_steps_per_second": 1.81,
|
|
"step": 385
|
|
},
|
|
{
|
|
"epoch": 10.13,
|
|
"learning_rate": 4.304016620498615e-05,
|
|
"loss": 0.0741,
|
|
"step": 390
|
|
},
|
|
{
|
|
"epoch": 10.39,
|
|
"learning_rate": 4.265927977839336e-05,
|
|
"loss": 0.0542,
|
|
"step": 400
|
|
},
|
|
{
|
|
"epoch": 10.65,
|
|
"learning_rate": 4.227839335180056e-05,
|
|
"loss": 0.0679,
|
|
"step": 410
|
|
},
|
|
{
|
|
"epoch": 10.91,
|
|
"learning_rate": 4.189750692520776e-05,
|
|
"loss": 0.0771,
|
|
"step": 420
|
|
},
|
|
{
|
|
"epoch": 10.99,
|
|
"eval_accuracy": 0.7666666666666667,
|
|
"eval_loss": 0.9409402012825012,
|
|
"eval_runtime": 1.2296,
|
|
"eval_samples_per_second": 48.796,
|
|
"eval_steps_per_second": 1.627,
|
|
"step": 423
|
|
},
|
|
{
|
|
"epoch": 11.17,
|
|
"learning_rate": 4.151662049861496e-05,
|
|
"loss": 0.0665,
|
|
"step": 430
|
|
},
|
|
{
|
|
"epoch": 11.43,
|
|
"learning_rate": 4.113573407202216e-05,
|
|
"loss": 0.0413,
|
|
"step": 440
|
|
},
|
|
{
|
|
"epoch": 11.69,
|
|
"learning_rate": 4.075484764542937e-05,
|
|
"loss": 0.0644,
|
|
"step": 450
|
|
},
|
|
{
|
|
"epoch": 11.95,
|
|
"learning_rate": 4.037396121883656e-05,
|
|
"loss": 0.0527,
|
|
"step": 460
|
|
},
|
|
{
|
|
"epoch": 12.0,
|
|
"eval_accuracy": 0.7833333333333333,
|
|
"eval_loss": 1.0052143335342407,
|
|
"eval_runtime": 1.0721,
|
|
"eval_samples_per_second": 55.965,
|
|
"eval_steps_per_second": 1.865,
|
|
"step": 462
|
|
},
|
|
{
|
|
"epoch": 12.21,
|
|
"learning_rate": 3.999307479224377e-05,
|
|
"loss": 0.064,
|
|
"step": 470
|
|
},
|
|
{
|
|
"epoch": 12.47,
|
|
"learning_rate": 3.961218836565097e-05,
|
|
"loss": 0.0562,
|
|
"step": 480
|
|
},
|
|
{
|
|
"epoch": 12.73,
|
|
"learning_rate": 3.923130193905817e-05,
|
|
"loss": 0.0678,
|
|
"step": 490
|
|
},
|
|
{
|
|
"epoch": 12.99,
|
|
"learning_rate": 3.885041551246538e-05,
|
|
"loss": 0.0708,
|
|
"step": 500
|
|
},
|
|
{
|
|
"epoch": 12.99,
|
|
"eval_accuracy": 0.8,
|
|
"eval_loss": 0.9578073024749756,
|
|
"eval_runtime": 1.0768,
|
|
"eval_samples_per_second": 55.72,
|
|
"eval_steps_per_second": 1.857,
|
|
"step": 500
|
|
},
|
|
{
|
|
"epoch": 13.25,
|
|
"learning_rate": 3.8469529085872574e-05,
|
|
"loss": 0.0429,
|
|
"step": 510
|
|
},
|
|
{
|
|
"epoch": 13.51,
|
|
"learning_rate": 3.808864265927978e-05,
|
|
"loss": 0.0553,
|
|
"step": 520
|
|
},
|
|
{
|
|
"epoch": 13.77,
|
|
"learning_rate": 3.770775623268698e-05,
|
|
"loss": 0.0562,
|
|
"step": 530
|
|
},
|
|
{
|
|
"epoch": 14.0,
|
|
"eval_accuracy": 0.8166666666666667,
|
|
"eval_loss": 1.0712392330169678,
|
|
"eval_runtime": 1.0829,
|
|
"eval_samples_per_second": 55.404,
|
|
"eval_steps_per_second": 1.847,
|
|
"step": 539
|
|
},
|
|
{
|
|
"epoch": 14.03,
|
|
"learning_rate": 3.732686980609418e-05,
|
|
"loss": 0.0448,
|
|
"step": 540
|
|
},
|
|
{
|
|
"epoch": 14.29,
|
|
"learning_rate": 3.694598337950139e-05,
|
|
"loss": 0.041,
|
|
"step": 550
|
|
},
|
|
{
|
|
"epoch": 14.55,
|
|
"learning_rate": 3.6565096952908585e-05,
|
|
"loss": 0.0577,
|
|
"step": 560
|
|
},
|
|
{
|
|
"epoch": 14.81,
|
|
"learning_rate": 3.618421052631579e-05,
|
|
"loss": 0.0467,
|
|
"step": 570
|
|
},
|
|
{
|
|
"epoch": 14.99,
|
|
"eval_accuracy": 0.8166666666666667,
|
|
"eval_loss": 1.058632731437683,
|
|
"eval_runtime": 1.2825,
|
|
"eval_samples_per_second": 46.783,
|
|
"eval_steps_per_second": 1.559,
|
|
"step": 577
|
|
},
|
|
{
|
|
"epoch": 15.06,
|
|
"learning_rate": 3.5803324099722994e-05,
|
|
"loss": 0.035,
|
|
"step": 580
|
|
},
|
|
{
|
|
"epoch": 15.32,
|
|
"learning_rate": 3.5422437673130195e-05,
|
|
"loss": 0.0572,
|
|
"step": 590
|
|
},
|
|
{
|
|
"epoch": 15.58,
|
|
"learning_rate": 3.5041551246537395e-05,
|
|
"loss": 0.0451,
|
|
"step": 600
|
|
},
|
|
{
|
|
"epoch": 15.84,
|
|
"learning_rate": 3.4660664819944596e-05,
|
|
"loss": 0.0445,
|
|
"step": 610
|
|
},
|
|
{
|
|
"epoch": 16.0,
|
|
"eval_accuracy": 0.7666666666666667,
|
|
"eval_loss": 1.20657479763031,
|
|
"eval_runtime": 1.2453,
|
|
"eval_samples_per_second": 48.181,
|
|
"eval_steps_per_second": 1.606,
|
|
"step": 616
|
|
},
|
|
{
|
|
"epoch": 16.1,
|
|
"learning_rate": 3.4279778393351804e-05,
|
|
"loss": 0.0338,
|
|
"step": 620
|
|
},
|
|
{
|
|
"epoch": 16.36,
|
|
"learning_rate": 3.3898891966759005e-05,
|
|
"loss": 0.0448,
|
|
"step": 630
|
|
},
|
|
{
|
|
"epoch": 16.62,
|
|
"learning_rate": 3.3518005540166206e-05,
|
|
"loss": 0.0488,
|
|
"step": 640
|
|
},
|
|
{
|
|
"epoch": 16.88,
|
|
"learning_rate": 3.313711911357341e-05,
|
|
"loss": 0.0474,
|
|
"step": 650
|
|
},
|
|
{
|
|
"epoch": 16.99,
|
|
"eval_accuracy": 0.75,
|
|
"eval_loss": 1.1863175630569458,
|
|
"eval_runtime": 1.0935,
|
|
"eval_samples_per_second": 54.87,
|
|
"eval_steps_per_second": 1.829,
|
|
"step": 654
|
|
},
|
|
{
|
|
"epoch": 17.14,
|
|
"learning_rate": 3.275623268698061e-05,
|
|
"loss": 0.0389,
|
|
"step": 660
|
|
},
|
|
{
|
|
"epoch": 17.4,
|
|
"learning_rate": 3.2375346260387815e-05,
|
|
"loss": 0.0442,
|
|
"step": 670
|
|
},
|
|
{
|
|
"epoch": 17.66,
|
|
"learning_rate": 3.1994459833795016e-05,
|
|
"loss": 0.0388,
|
|
"step": 680
|
|
},
|
|
{
|
|
"epoch": 17.92,
|
|
"learning_rate": 3.161357340720222e-05,
|
|
"loss": 0.0263,
|
|
"step": 690
|
|
},
|
|
{
|
|
"epoch": 18.0,
|
|
"eval_accuracy": 0.8166666666666667,
|
|
"eval_loss": 1.1207224130630493,
|
|
"eval_runtime": 1.0882,
|
|
"eval_samples_per_second": 55.137,
|
|
"eval_steps_per_second": 1.838,
|
|
"step": 693
|
|
},
|
|
{
|
|
"epoch": 18.18,
|
|
"learning_rate": 3.123268698060942e-05,
|
|
"loss": 0.0379,
|
|
"step": 700
|
|
},
|
|
{
|
|
"epoch": 18.44,
|
|
"learning_rate": 3.085180055401662e-05,
|
|
"loss": 0.0614,
|
|
"step": 710
|
|
},
|
|
{
|
|
"epoch": 18.7,
|
|
"learning_rate": 3.0470914127423827e-05,
|
|
"loss": 0.0367,
|
|
"step": 720
|
|
},
|
|
{
|
|
"epoch": 18.96,
|
|
"learning_rate": 3.0090027700831024e-05,
|
|
"loss": 0.0307,
|
|
"step": 730
|
|
},
|
|
{
|
|
"epoch": 18.99,
|
|
"eval_accuracy": 0.8166666666666667,
|
|
"eval_loss": 1.1812899112701416,
|
|
"eval_runtime": 1.0999,
|
|
"eval_samples_per_second": 54.549,
|
|
"eval_steps_per_second": 1.818,
|
|
"step": 731
|
|
},
|
|
{
|
|
"epoch": 19.22,
|
|
"learning_rate": 2.9709141274238225e-05,
|
|
"loss": 0.041,
|
|
"step": 740
|
|
},
|
|
{
|
|
"epoch": 19.48,
|
|
"learning_rate": 2.9328254847645433e-05,
|
|
"loss": 0.0367,
|
|
"step": 750
|
|
},
|
|
{
|
|
"epoch": 19.74,
|
|
"learning_rate": 2.894736842105263e-05,
|
|
"loss": 0.0311,
|
|
"step": 760
|
|
},
|
|
{
|
|
"epoch": 20.0,
|
|
"learning_rate": 2.8566481994459838e-05,
|
|
"loss": 0.0393,
|
|
"step": 770
|
|
},
|
|
{
|
|
"epoch": 20.0,
|
|
"eval_accuracy": 0.75,
|
|
"eval_loss": 1.3761276006698608,
|
|
"eval_runtime": 1.1025,
|
|
"eval_samples_per_second": 54.423,
|
|
"eval_steps_per_second": 1.814,
|
|
"step": 770
|
|
},
|
|
{
|
|
"epoch": 20.26,
|
|
"learning_rate": 2.8185595567867035e-05,
|
|
"loss": 0.0213,
|
|
"step": 780
|
|
},
|
|
{
|
|
"epoch": 20.52,
|
|
"learning_rate": 2.7804709141274236e-05,
|
|
"loss": 0.0206,
|
|
"step": 790
|
|
},
|
|
{
|
|
"epoch": 20.78,
|
|
"learning_rate": 2.742382271468144e-05,
|
|
"loss": 0.0475,
|
|
"step": 800
|
|
},
|
|
{
|
|
"epoch": 20.99,
|
|
"eval_accuracy": 0.7666666666666667,
|
|
"eval_loss": 1.3007997274398804,
|
|
"eval_runtime": 1.252,
|
|
"eval_samples_per_second": 47.924,
|
|
"eval_steps_per_second": 1.597,
|
|
"step": 808
|
|
},
|
|
{
|
|
"epoch": 21.04,
|
|
"learning_rate": 2.7042936288088645e-05,
|
|
"loss": 0.0447,
|
|
"step": 810
|
|
},
|
|
{
|
|
"epoch": 21.3,
|
|
"learning_rate": 2.6662049861495846e-05,
|
|
"loss": 0.0274,
|
|
"step": 820
|
|
},
|
|
{
|
|
"epoch": 21.56,
|
|
"learning_rate": 2.6281163434903047e-05,
|
|
"loss": 0.0369,
|
|
"step": 830
|
|
},
|
|
{
|
|
"epoch": 21.82,
|
|
"learning_rate": 2.590027700831025e-05,
|
|
"loss": 0.0215,
|
|
"step": 840
|
|
},
|
|
{
|
|
"epoch": 22.0,
|
|
"eval_accuracy": 0.7333333333333333,
|
|
"eval_loss": 1.2625234127044678,
|
|
"eval_runtime": 1.2982,
|
|
"eval_samples_per_second": 46.219,
|
|
"eval_steps_per_second": 1.541,
|
|
"step": 847
|
|
},
|
|
{
|
|
"epoch": 22.08,
|
|
"learning_rate": 2.5519390581717452e-05,
|
|
"loss": 0.0434,
|
|
"step": 850
|
|
},
|
|
{
|
|
"epoch": 22.34,
|
|
"learning_rate": 2.5138504155124656e-05,
|
|
"loss": 0.0338,
|
|
"step": 860
|
|
},
|
|
{
|
|
"epoch": 22.6,
|
|
"learning_rate": 2.4757617728531857e-05,
|
|
"loss": 0.0275,
|
|
"step": 870
|
|
},
|
|
{
|
|
"epoch": 22.86,
|
|
"learning_rate": 2.4376731301939058e-05,
|
|
"loss": 0.0311,
|
|
"step": 880
|
|
},
|
|
{
|
|
"epoch": 22.99,
|
|
"eval_accuracy": 0.8,
|
|
"eval_loss": 1.1508437395095825,
|
|
"eval_runtime": 0.9933,
|
|
"eval_samples_per_second": 60.407,
|
|
"eval_steps_per_second": 2.014,
|
|
"step": 885
|
|
},
|
|
{
|
|
"epoch": 23.12,
|
|
"learning_rate": 2.3995844875346262e-05,
|
|
"loss": 0.0332,
|
|
"step": 890
|
|
},
|
|
{
|
|
"epoch": 23.38,
|
|
"learning_rate": 2.3614958448753463e-05,
|
|
"loss": 0.0503,
|
|
"step": 900
|
|
},
|
|
{
|
|
"epoch": 23.64,
|
|
"learning_rate": 2.3234072022160667e-05,
|
|
"loss": 0.0267,
|
|
"step": 910
|
|
},
|
|
{
|
|
"epoch": 23.9,
|
|
"learning_rate": 2.2853185595567865e-05,
|
|
"loss": 0.027,
|
|
"step": 920
|
|
},
|
|
{
|
|
"epoch": 24.0,
|
|
"eval_accuracy": 0.7666666666666667,
|
|
"eval_loss": 1.3034749031066895,
|
|
"eval_runtime": 1.0568,
|
|
"eval_samples_per_second": 56.773,
|
|
"eval_steps_per_second": 1.892,
|
|
"step": 924
|
|
},
|
|
{
|
|
"epoch": 24.16,
|
|
"learning_rate": 2.247229916897507e-05,
|
|
"loss": 0.0309,
|
|
"step": 930
|
|
},
|
|
{
|
|
"epoch": 24.42,
|
|
"learning_rate": 2.2091412742382273e-05,
|
|
"loss": 0.0194,
|
|
"step": 940
|
|
},
|
|
{
|
|
"epoch": 24.68,
|
|
"learning_rate": 2.1710526315789474e-05,
|
|
"loss": 0.0296,
|
|
"step": 950
|
|
},
|
|
{
|
|
"epoch": 24.94,
|
|
"learning_rate": 2.132963988919668e-05,
|
|
"loss": 0.0251,
|
|
"step": 960
|
|
},
|
|
{
|
|
"epoch": 24.99,
|
|
"eval_accuracy": 0.7666666666666667,
|
|
"eval_loss": 1.2270464897155762,
|
|
"eval_runtime": 1.0482,
|
|
"eval_samples_per_second": 57.24,
|
|
"eval_steps_per_second": 1.908,
|
|
"step": 962
|
|
},
|
|
{
|
|
"epoch": 25.19,
|
|
"learning_rate": 2.094875346260388e-05,
|
|
"loss": 0.0262,
|
|
"step": 970
|
|
},
|
|
{
|
|
"epoch": 25.45,
|
|
"learning_rate": 2.056786703601108e-05,
|
|
"loss": 0.0253,
|
|
"step": 980
|
|
},
|
|
{
|
|
"epoch": 25.71,
|
|
"learning_rate": 2.018698060941828e-05,
|
|
"loss": 0.0196,
|
|
"step": 990
|
|
},
|
|
{
|
|
"epoch": 25.97,
|
|
"learning_rate": 1.9806094182825486e-05,
|
|
"loss": 0.0161,
|
|
"step": 1000
|
|
},
|
|
{
|
|
"epoch": 26.0,
|
|
"eval_accuracy": 0.8166666666666667,
|
|
"eval_loss": 1.1470165252685547,
|
|
"eval_runtime": 1.06,
|
|
"eval_samples_per_second": 56.602,
|
|
"eval_steps_per_second": 1.887,
|
|
"step": 1001
|
|
},
|
|
{
|
|
"epoch": 26.23,
|
|
"learning_rate": 1.942520775623269e-05,
|
|
"loss": 0.027,
|
|
"step": 1010
|
|
},
|
|
{
|
|
"epoch": 26.49,
|
|
"learning_rate": 1.904432132963989e-05,
|
|
"loss": 0.0262,
|
|
"step": 1020
|
|
},
|
|
{
|
|
"epoch": 26.75,
|
|
"learning_rate": 1.866343490304709e-05,
|
|
"loss": 0.0258,
|
|
"step": 1030
|
|
},
|
|
{
|
|
"epoch": 26.99,
|
|
"eval_accuracy": 0.8166666666666667,
|
|
"eval_loss": 1.1472843885421753,
|
|
"eval_runtime": 1.07,
|
|
"eval_samples_per_second": 56.076,
|
|
"eval_steps_per_second": 1.869,
|
|
"step": 1039
|
|
},
|
|
{
|
|
"epoch": 27.01,
|
|
"learning_rate": 1.8282548476454293e-05,
|
|
"loss": 0.0192,
|
|
"step": 1040
|
|
},
|
|
{
|
|
"epoch": 27.27,
|
|
"learning_rate": 1.7901662049861497e-05,
|
|
"loss": 0.0165,
|
|
"step": 1050
|
|
},
|
|
{
|
|
"epoch": 27.53,
|
|
"learning_rate": 1.7520775623268698e-05,
|
|
"loss": 0.0209,
|
|
"step": 1060
|
|
},
|
|
{
|
|
"epoch": 27.79,
|
|
"learning_rate": 1.7139889196675902e-05,
|
|
"loss": 0.0142,
|
|
"step": 1070
|
|
},
|
|
{
|
|
"epoch": 28.0,
|
|
"eval_accuracy": 0.7666666666666667,
|
|
"eval_loss": 1.2325979471206665,
|
|
"eval_runtime": 1.1661,
|
|
"eval_samples_per_second": 51.455,
|
|
"eval_steps_per_second": 1.715,
|
|
"step": 1078
|
|
},
|
|
{
|
|
"epoch": 28.05,
|
|
"learning_rate": 1.6759002770083103e-05,
|
|
"loss": 0.0178,
|
|
"step": 1080
|
|
},
|
|
{
|
|
"epoch": 28.31,
|
|
"learning_rate": 1.6378116343490304e-05,
|
|
"loss": 0.0214,
|
|
"step": 1090
|
|
},
|
|
{
|
|
"epoch": 28.57,
|
|
"learning_rate": 1.5997229916897508e-05,
|
|
"loss": 0.0236,
|
|
"step": 1100
|
|
},
|
|
{
|
|
"epoch": 28.83,
|
|
"learning_rate": 1.561634349030471e-05,
|
|
"loss": 0.0151,
|
|
"step": 1110
|
|
},
|
|
{
|
|
"epoch": 28.99,
|
|
"eval_accuracy": 0.7666666666666667,
|
|
"eval_loss": 1.3977586030960083,
|
|
"eval_runtime": 1.0743,
|
|
"eval_samples_per_second": 55.848,
|
|
"eval_steps_per_second": 1.862,
|
|
"step": 1116
|
|
},
|
|
{
|
|
"epoch": 29.09,
|
|
"learning_rate": 1.5235457063711913e-05,
|
|
"loss": 0.0259,
|
|
"step": 1120
|
|
},
|
|
{
|
|
"epoch": 29.35,
|
|
"learning_rate": 1.4854570637119112e-05,
|
|
"loss": 0.0181,
|
|
"step": 1130
|
|
},
|
|
{
|
|
"epoch": 29.61,
|
|
"learning_rate": 1.4473684210526315e-05,
|
|
"loss": 0.0157,
|
|
"step": 1140
|
|
},
|
|
{
|
|
"epoch": 29.87,
|
|
"learning_rate": 1.4092797783933518e-05,
|
|
"loss": 0.021,
|
|
"step": 1150
|
|
},
|
|
{
|
|
"epoch": 30.0,
|
|
"eval_accuracy": 0.8333333333333334,
|
|
"eval_loss": 1.2002999782562256,
|
|
"eval_runtime": 1.0917,
|
|
"eval_samples_per_second": 54.958,
|
|
"eval_steps_per_second": 1.832,
|
|
"step": 1155
|
|
},
|
|
{
|
|
"epoch": 30.13,
|
|
"learning_rate": 1.371191135734072e-05,
|
|
"loss": 0.0184,
|
|
"step": 1160
|
|
},
|
|
{
|
|
"epoch": 30.39,
|
|
"learning_rate": 1.3331024930747923e-05,
|
|
"loss": 0.0163,
|
|
"step": 1170
|
|
},
|
|
{
|
|
"epoch": 30.65,
|
|
"learning_rate": 1.2950138504155125e-05,
|
|
"loss": 0.0254,
|
|
"step": 1180
|
|
},
|
|
{
|
|
"epoch": 30.91,
|
|
"learning_rate": 1.2569252077562328e-05,
|
|
"loss": 0.0158,
|
|
"step": 1190
|
|
},
|
|
{
|
|
"epoch": 30.99,
|
|
"eval_accuracy": 0.7666666666666667,
|
|
"eval_loss": 1.2488106489181519,
|
|
"eval_runtime": 1.044,
|
|
"eval_samples_per_second": 57.472,
|
|
"eval_steps_per_second": 1.916,
|
|
"step": 1193
|
|
},
|
|
{
|
|
"epoch": 31.17,
|
|
"learning_rate": 1.2188365650969529e-05,
|
|
"loss": 0.0232,
|
|
"step": 1200
|
|
},
|
|
{
|
|
"epoch": 31.43,
|
|
"learning_rate": 1.1807479224376732e-05,
|
|
"loss": 0.0099,
|
|
"step": 1210
|
|
},
|
|
{
|
|
"epoch": 31.69,
|
|
"learning_rate": 1.1426592797783932e-05,
|
|
"loss": 0.0211,
|
|
"step": 1220
|
|
},
|
|
{
|
|
"epoch": 31.95,
|
|
"learning_rate": 1.1045706371191137e-05,
|
|
"loss": 0.0163,
|
|
"step": 1230
|
|
},
|
|
{
|
|
"epoch": 32.0,
|
|
"eval_accuracy": 0.75,
|
|
"eval_loss": 1.3231933116912842,
|
|
"eval_runtime": 1.0878,
|
|
"eval_samples_per_second": 55.156,
|
|
"eval_steps_per_second": 1.839,
|
|
"step": 1232
|
|
},
|
|
{
|
|
"epoch": 32.21,
|
|
"learning_rate": 1.066481994459834e-05,
|
|
"loss": 0.0127,
|
|
"step": 1240
|
|
},
|
|
{
|
|
"epoch": 32.47,
|
|
"learning_rate": 1.028393351800554e-05,
|
|
"loss": 0.0206,
|
|
"step": 1250
|
|
},
|
|
{
|
|
"epoch": 32.73,
|
|
"learning_rate": 9.903047091412743e-06,
|
|
"loss": 0.0232,
|
|
"step": 1260
|
|
},
|
|
{
|
|
"epoch": 32.99,
|
|
"learning_rate": 9.522160664819945e-06,
|
|
"loss": 0.0143,
|
|
"step": 1270
|
|
},
|
|
{
|
|
"epoch": 32.99,
|
|
"eval_accuracy": 0.8,
|
|
"eval_loss": 1.246673583984375,
|
|
"eval_runtime": 1.0784,
|
|
"eval_samples_per_second": 55.637,
|
|
"eval_steps_per_second": 1.855,
|
|
"step": 1270
|
|
},
|
|
{
|
|
"epoch": 33.25,
|
|
"learning_rate": 9.141274238227146e-06,
|
|
"loss": 0.009,
|
|
"step": 1280
|
|
},
|
|
{
|
|
"epoch": 33.51,
|
|
"learning_rate": 8.760387811634349e-06,
|
|
"loss": 0.0208,
|
|
"step": 1290
|
|
},
|
|
{
|
|
"epoch": 33.77,
|
|
"learning_rate": 8.379501385041551e-06,
|
|
"loss": 0.02,
|
|
"step": 1300
|
|
},
|
|
{
|
|
"epoch": 34.0,
|
|
"eval_accuracy": 0.7833333333333333,
|
|
"eval_loss": 1.3175911903381348,
|
|
"eval_runtime": 1.0636,
|
|
"eval_samples_per_second": 56.415,
|
|
"eval_steps_per_second": 1.88,
|
|
"step": 1309
|
|
},
|
|
{
|
|
"epoch": 34.03,
|
|
"learning_rate": 7.998614958448754e-06,
|
|
"loss": 0.0179,
|
|
"step": 1310
|
|
},
|
|
{
|
|
"epoch": 34.29,
|
|
"learning_rate": 7.617728531855957e-06,
|
|
"loss": 0.0148,
|
|
"step": 1320
|
|
},
|
|
{
|
|
"epoch": 34.55,
|
|
"learning_rate": 7.2368421052631575e-06,
|
|
"loss": 0.0168,
|
|
"step": 1330
|
|
},
|
|
{
|
|
"epoch": 34.81,
|
|
"learning_rate": 6.85595567867036e-06,
|
|
"loss": 0.0128,
|
|
"step": 1340
|
|
},
|
|
{
|
|
"epoch": 34.99,
|
|
"eval_accuracy": 0.7666666666666667,
|
|
"eval_loss": 1.3082655668258667,
|
|
"eval_runtime": 1.0769,
|
|
"eval_samples_per_second": 55.716,
|
|
"eval_steps_per_second": 1.857,
|
|
"step": 1347
|
|
},
|
|
{
|
|
"epoch": 35.06,
|
|
"learning_rate": 6.475069252077563e-06,
|
|
"loss": 0.018,
|
|
"step": 1350
|
|
},
|
|
{
|
|
"epoch": 35.32,
|
|
"learning_rate": 6.0941828254847645e-06,
|
|
"loss": 0.0123,
|
|
"step": 1360
|
|
},
|
|
{
|
|
"epoch": 35.58,
|
|
"learning_rate": 5.713296398891966e-06,
|
|
"loss": 0.0124,
|
|
"step": 1370
|
|
},
|
|
{
|
|
"epoch": 35.84,
|
|
"learning_rate": 5.33240997229917e-06,
|
|
"loss": 0.0144,
|
|
"step": 1380
|
|
},
|
|
{
|
|
"epoch": 36.0,
|
|
"eval_accuracy": 0.7666666666666667,
|
|
"eval_loss": 1.3079514503479004,
|
|
"eval_runtime": 1.0808,
|
|
"eval_samples_per_second": 55.512,
|
|
"eval_steps_per_second": 1.85,
|
|
"step": 1386
|
|
},
|
|
{
|
|
"epoch": 36.1,
|
|
"learning_rate": 4.951523545706371e-06,
|
|
"loss": 0.0156,
|
|
"step": 1390
|
|
},
|
|
{
|
|
"epoch": 36.36,
|
|
"learning_rate": 4.570637119113573e-06,
|
|
"loss": 0.0124,
|
|
"step": 1400
|
|
},
|
|
{
|
|
"epoch": 36.62,
|
|
"learning_rate": 4.189750692520776e-06,
|
|
"loss": 0.0134,
|
|
"step": 1410
|
|
},
|
|
{
|
|
"epoch": 36.88,
|
|
"learning_rate": 3.8088642659279783e-06,
|
|
"loss": 0.0109,
|
|
"step": 1420
|
|
},
|
|
{
|
|
"epoch": 36.99,
|
|
"eval_accuracy": 0.8,
|
|
"eval_loss": 1.2998783588409424,
|
|
"eval_runtime": 1.0406,
|
|
"eval_samples_per_second": 57.661,
|
|
"eval_steps_per_second": 1.922,
|
|
"step": 1424
|
|
},
|
|
{
|
|
"epoch": 37.14,
|
|
"learning_rate": 3.42797783933518e-06,
|
|
"loss": 0.0147,
|
|
"step": 1430
|
|
},
|
|
{
|
|
"epoch": 37.4,
|
|
"learning_rate": 3.0470914127423822e-06,
|
|
"loss": 0.0116,
|
|
"step": 1440
|
|
},
|
|
{
|
|
"epoch": 37.66,
|
|
"learning_rate": 2.666204986149585e-06,
|
|
"loss": 0.0102,
|
|
"step": 1450
|
|
},
|
|
{
|
|
"epoch": 37.92,
|
|
"learning_rate": 2.2853185595567866e-06,
|
|
"loss": 0.0082,
|
|
"step": 1460
|
|
},
|
|
{
|
|
"epoch": 38.0,
|
|
"eval_accuracy": 0.8,
|
|
"eval_loss": 1.2717970609664917,
|
|
"eval_runtime": 1.0837,
|
|
"eval_samples_per_second": 55.367,
|
|
"eval_steps_per_second": 1.846,
|
|
"step": 1463
|
|
},
|
|
{
|
|
"epoch": 38.18,
|
|
"learning_rate": 1.9044321329639892e-06,
|
|
"loss": 0.0103,
|
|
"step": 1470
|
|
},
|
|
{
|
|
"epoch": 38.44,
|
|
"learning_rate": 1.5235457063711911e-06,
|
|
"loss": 0.0207,
|
|
"step": 1480
|
|
},
|
|
{
|
|
"epoch": 38.7,
|
|
"learning_rate": 1.1426592797783933e-06,
|
|
"loss": 0.0098,
|
|
"step": 1490
|
|
},
|
|
{
|
|
"epoch": 38.96,
|
|
"learning_rate": 7.617728531855956e-07,
|
|
"loss": 0.0064,
|
|
"step": 1500
|
|
},
|
|
{
|
|
"epoch": 38.99,
|
|
"eval_accuracy": 0.7666666666666667,
|
|
"eval_loss": 1.2587716579437256,
|
|
"eval_runtime": 1.1594,
|
|
"eval_samples_per_second": 51.749,
|
|
"eval_steps_per_second": 1.725,
|
|
"step": 1501
|
|
},
|
|
{
|
|
"epoch": 39.22,
|
|
"learning_rate": 3.808864265927978e-07,
|
|
"loss": 0.0137,
|
|
"step": 1510
|
|
},
|
|
{
|
|
"epoch": 39.48,
|
|
"learning_rate": 0.0,
|
|
"loss": 0.0097,
|
|
"step": 1520
|
|
},
|
|
{
|
|
"epoch": 39.48,
|
|
"eval_accuracy": 0.7666666666666667,
|
|
"eval_loss": 1.2597320079803467,
|
|
"eval_runtime": 1.105,
|
|
"eval_samples_per_second": 54.301,
|
|
"eval_steps_per_second": 1.81,
|
|
"step": 1520
|
|
},
|
|
{
|
|
"epoch": 39.48,
|
|
"step": 1520,
|
|
"total_flos": 1.4992488987811873e+19,
|
|
"train_loss": 0.11935894162835259,
|
|
"train_runtime": 3007.0998,
|
|
"train_samples_per_second": 65.179,
|
|
"train_steps_per_second": 0.505
|
|
}
|
|
],
|
|
"logging_steps": 10,
|
|
"max_steps": 1520,
|
|
"num_input_tokens_seen": 0,
|
|
"num_train_epochs": 40,
|
|
"save_steps": 500,
|
|
"total_flos": 1.4992488987811873e+19,
|
|
"train_batch_size": 32,
|
|
"trial_name": null,
|
|
"trial_params": null
|
|
}
|
|
|