{ "best_metric": 29.65376782077393, "best_model_checkpoint": "./checkpoint-4000", "epoch": 92.0064, "eval_steps": 500, "global_step": 5000, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.01, "learning_rate": 4.800000000000001e-07, "loss": 1.4417, "step": 25 }, { "epoch": 0.01, "learning_rate": 9.800000000000001e-07, "loss": 1.0991, "step": 50 }, { "epoch": 1.0, "learning_rate": 1.48e-06, "loss": 0.6718, "step": 75 }, { "epoch": 1.01, "learning_rate": 1.98e-06, "loss": 0.5142, "step": 100 }, { "epoch": 2.0, "learning_rate": 2.4800000000000004e-06, "loss": 0.3715, "step": 125 }, { "epoch": 2.01, "learning_rate": 2.9800000000000003e-06, "loss": 0.3011, "step": 150 }, { "epoch": 3.0, "learning_rate": 3.48e-06, "loss": 0.2129, "step": 175 }, { "epoch": 3.01, "learning_rate": 3.980000000000001e-06, "loss": 0.1469, "step": 200 }, { "epoch": 4.0, "learning_rate": 4.48e-06, "loss": 0.1142, "step": 225 }, { "epoch": 4.01, "learning_rate": 4.980000000000001e-06, "loss": 0.0612, "step": 250 }, { "epoch": 5.0, "learning_rate": 5.480000000000001e-06, "loss": 0.076, "step": 275 }, { "epoch": 5.01, "learning_rate": 5.98e-06, "loss": 0.0364, "step": 300 }, { "epoch": 6.0, "learning_rate": 6.480000000000001e-06, "loss": 0.0462, "step": 325 }, { "epoch": 6.01, "learning_rate": 6.98e-06, "loss": 0.0313, "step": 350 }, { "epoch": 6.01, "learning_rate": 7.48e-06, "loss": 0.0375, "step": 375 }, { "epoch": 7.0, "learning_rate": 7.980000000000002e-06, "loss": 0.0262, "step": 400 }, { "epoch": 7.01, "learning_rate": 8.48e-06, "loss": 0.031, "step": 425 }, { "epoch": 8.0, "learning_rate": 8.98e-06, "loss": 0.0331, "step": 450 }, { "epoch": 8.01, "learning_rate": 9.48e-06, "loss": 0.0373, "step": 475 }, { "epoch": 9.0, "learning_rate": 9.980000000000001e-06, "loss": 0.0341, "step": 500 }, { "epoch": 9.0, "eval_loss": 0.545076310634613, "eval_runtime": 109.8569, "eval_samples_per_second": 3.495, "eval_steps_per_second": 0.218, "eval_wer": 34.74541751527495, "step": 500 }, { "epoch": 9.01, "learning_rate": 9.946666666666667e-06, "loss": 0.03, "step": 525 }, { "epoch": 10.0, "learning_rate": 9.893333333333334e-06, "loss": 0.0347, "step": 550 }, { "epoch": 10.01, "learning_rate": 9.837777777777778e-06, "loss": 0.0273, "step": 575 }, { "epoch": 11.0, "learning_rate": 9.782222222222222e-06, "loss": 0.0251, "step": 600 }, { "epoch": 11.01, "learning_rate": 9.726666666666668e-06, "loss": 0.0207, "step": 625 }, { "epoch": 12.0, "learning_rate": 9.671111111111112e-06, "loss": 0.0189, "step": 650 }, { "epoch": 12.01, "learning_rate": 9.615555555555558e-06, "loss": 0.016, "step": 675 }, { "epoch": 12.01, "learning_rate": 9.56e-06, "loss": 0.019, "step": 700 }, { "epoch": 13.0, "learning_rate": 9.504444444444446e-06, "loss": 0.0156, "step": 725 }, { "epoch": 13.01, "learning_rate": 9.44888888888889e-06, "loss": 0.0101, "step": 750 }, { "epoch": 14.0, "learning_rate": 9.393333333333334e-06, "loss": 0.014, "step": 775 }, { "epoch": 14.01, "learning_rate": 9.33777777777778e-06, "loss": 0.0109, "step": 800 }, { "epoch": 15.0, "learning_rate": 9.282222222222222e-06, "loss": 0.0096, "step": 825 }, { "epoch": 15.01, "learning_rate": 9.226666666666668e-06, "loss": 0.0193, "step": 850 }, { "epoch": 16.0, "learning_rate": 9.171111111111112e-06, "loss": 0.0077, "step": 875 }, { "epoch": 16.01, "learning_rate": 9.115555555555556e-06, "loss": 0.0143, "step": 900 }, { "epoch": 17.0, "learning_rate": 9.060000000000001e-06, "loss": 0.0123, "step": 925 }, { "epoch": 17.01, "learning_rate": 9.004444444444445e-06, "loss": 0.0149, "step": 950 }, { "epoch": 18.0, "learning_rate": 8.94888888888889e-06, "loss": 0.0129, "step": 975 }, { "epoch": 18.01, "learning_rate": 8.893333333333333e-06, "loss": 0.0107, "step": 1000 }, { "epoch": 18.01, "eval_loss": 0.5564951300621033, "eval_runtime": 121.1746, "eval_samples_per_second": 3.169, "eval_steps_per_second": 0.198, "eval_wer": 31.486761710794298, "step": 1000 }, { "epoch": 18.01, "learning_rate": 8.83777777777778e-06, "loss": 0.007, "step": 1025 }, { "epoch": 19.0, "learning_rate": 8.782222222222223e-06, "loss": 0.0149, "step": 1050 }, { "epoch": 19.01, "learning_rate": 8.726666666666667e-06, "loss": 0.0047, "step": 1075 }, { "epoch": 20.0, "learning_rate": 8.671111111111113e-06, "loss": 0.0106, "step": 1100 }, { "epoch": 20.01, "learning_rate": 8.615555555555555e-06, "loss": 0.0061, "step": 1125 }, { "epoch": 21.0, "learning_rate": 8.560000000000001e-06, "loss": 0.005, "step": 1150 }, { "epoch": 21.01, "learning_rate": 8.504444444444445e-06, "loss": 0.0058, "step": 1175 }, { "epoch": 22.0, "learning_rate": 8.448888888888889e-06, "loss": 0.0042, "step": 1200 }, { "epoch": 22.01, "learning_rate": 8.393333333333335e-06, "loss": 0.0019, "step": 1225 }, { "epoch": 23.0, "learning_rate": 8.337777777777777e-06, "loss": 0.003, "step": 1250 }, { "epoch": 23.01, "learning_rate": 8.282222222222223e-06, "loss": 0.0036, "step": 1275 }, { "epoch": 24.0, "learning_rate": 8.226666666666667e-06, "loss": 0.001, "step": 1300 }, { "epoch": 24.01, "learning_rate": 8.171111111111113e-06, "loss": 0.0004, "step": 1325 }, { "epoch": 24.01, "learning_rate": 8.115555555555557e-06, "loss": 0.0044, "step": 1350 }, { "epoch": 25.0, "learning_rate": 8.06e-06, "loss": 0.0019, "step": 1375 }, { "epoch": 25.01, "learning_rate": 8.004444444444445e-06, "loss": 0.0017, "step": 1400 }, { "epoch": 26.0, "learning_rate": 7.948888888888889e-06, "loss": 0.0004, "step": 1425 }, { "epoch": 26.01, "learning_rate": 7.893333333333335e-06, "loss": 0.0002, "step": 1450 }, { "epoch": 27.0, "learning_rate": 7.837777777777779e-06, "loss": 0.0002, "step": 1475 }, { "epoch": 27.01, "learning_rate": 7.782222222222223e-06, "loss": 0.0012, "step": 1500 }, { "epoch": 27.01, "eval_loss": 0.5967185497283936, "eval_runtime": 104.4619, "eval_samples_per_second": 3.676, "eval_steps_per_second": 0.23, "eval_wer": 30.509164969450104, "step": 1500 }, { "epoch": 28.0, "learning_rate": 7.726666666666667e-06, "loss": 0.0007, "step": 1525 }, { "epoch": 28.01, "learning_rate": 7.67111111111111e-06, "loss": 0.0008, "step": 1550 }, { "epoch": 29.0, "learning_rate": 7.6155555555555564e-06, "loss": 0.0005, "step": 1575 }, { "epoch": 29.01, "learning_rate": 7.5600000000000005e-06, "loss": 0.0014, "step": 1600 }, { "epoch": 30.0, "learning_rate": 7.504444444444445e-06, "loss": 0.0002, "step": 1625 }, { "epoch": 30.01, "learning_rate": 7.44888888888889e-06, "loss": 0.0007, "step": 1650 }, { "epoch": 31.0, "learning_rate": 7.393333333333333e-06, "loss": 0.0004, "step": 1675 }, { "epoch": 31.01, "learning_rate": 7.337777777777778e-06, "loss": 0.0002, "step": 1700 }, { "epoch": 31.01, "learning_rate": 7.282222222222222e-06, "loss": 0.0004, "step": 1725 }, { "epoch": 32.0, "learning_rate": 7.226666666666667e-06, "loss": 0.0011, "step": 1750 }, { "epoch": 32.01, "learning_rate": 7.171111111111112e-06, "loss": 0.0011, "step": 1775 }, { "epoch": 33.0, "learning_rate": 7.115555555555557e-06, "loss": 0.0003, "step": 1800 }, { "epoch": 33.01, "learning_rate": 7.06e-06, "loss": 0.0019, "step": 1825 }, { "epoch": 34.0, "learning_rate": 7.004444444444445e-06, "loss": 0.0007, "step": 1850 }, { "epoch": 34.01, "learning_rate": 6.948888888888889e-06, "loss": 0.0003, "step": 1875 }, { "epoch": 35.0, "learning_rate": 6.893333333333334e-06, "loss": 0.0002, "step": 1900 }, { "epoch": 35.01, "learning_rate": 6.837777777777779e-06, "loss": 0.0001, "step": 1925 }, { "epoch": 36.0, "learning_rate": 6.782222222222222e-06, "loss": 0.0001, "step": 1950 }, { "epoch": 36.01, "learning_rate": 6.726666666666667e-06, "loss": 0.0001, "step": 1975 }, { "epoch": 37.0, "learning_rate": 6.671111111111112e-06, "loss": 0.0001, "step": 2000 }, { "epoch": 37.0, "eval_loss": 0.6311168670654297, "eval_runtime": 112.6757, "eval_samples_per_second": 3.408, "eval_steps_per_second": 0.213, "eval_wer": 30.712830957230143, "step": 2000 }, { "epoch": 37.01, "learning_rate": 6.615555555555556e-06, "loss": 0.0001, "step": 2025 }, { "epoch": 37.01, "learning_rate": 6.560000000000001e-06, "loss": 0.0001, "step": 2050 }, { "epoch": 38.0, "learning_rate": 6.504444444444446e-06, "loss": 0.0001, "step": 2075 }, { "epoch": 38.01, "learning_rate": 6.448888888888889e-06, "loss": 0.0001, "step": 2100 }, { "epoch": 39.0, "learning_rate": 6.393333333333334e-06, "loss": 0.0001, "step": 2125 }, { "epoch": 39.01, "learning_rate": 6.3377777777777786e-06, "loss": 0.0001, "step": 2150 }, { "epoch": 40.0, "learning_rate": 6.282222222222223e-06, "loss": 0.0001, "step": 2175 }, { "epoch": 40.01, "learning_rate": 6.2266666666666675e-06, "loss": 0.0001, "step": 2200 }, { "epoch": 41.0, "learning_rate": 6.171111111111112e-06, "loss": 0.0001, "step": 2225 }, { "epoch": 41.01, "learning_rate": 6.1155555555555555e-06, "loss": 0.0001, "step": 2250 }, { "epoch": 42.0, "learning_rate": 6.0600000000000004e-06, "loss": 0.0001, "step": 2275 }, { "epoch": 42.01, "learning_rate": 6.004444444444445e-06, "loss": 0.0001, "step": 2300 }, { "epoch": 43.0, "learning_rate": 5.948888888888889e-06, "loss": 0.0001, "step": 2325 }, { "epoch": 43.01, "learning_rate": 5.893333333333334e-06, "loss": 0.0001, "step": 2350 }, { "epoch": 43.01, "learning_rate": 5.837777777777777e-06, "loss": 0.0001, "step": 2375 }, { "epoch": 44.0, "learning_rate": 5.782222222222222e-06, "loss": 0.0001, "step": 2400 }, { "epoch": 44.01, "learning_rate": 5.726666666666667e-06, "loss": 0.0001, "step": 2425 }, { "epoch": 45.0, "learning_rate": 5.671111111111112e-06, "loss": 0.0001, "step": 2450 }, { "epoch": 45.01, "learning_rate": 5.615555555555556e-06, "loss": 0.0001, "step": 2475 }, { "epoch": 46.0, "learning_rate": 5.560000000000001e-06, "loss": 0.0001, "step": 2500 }, { "epoch": 46.0, "eval_loss": 0.6630290150642395, "eval_runtime": 105.6948, "eval_samples_per_second": 3.633, "eval_steps_per_second": 0.227, "eval_wer": 30.061099796334013, "step": 2500 }, { "epoch": 46.01, "learning_rate": 5.504444444444444e-06, "loss": 0.0001, "step": 2525 }, { "epoch": 47.0, "learning_rate": 5.448888888888889e-06, "loss": 0.0001, "step": 2550 }, { "epoch": 47.01, "learning_rate": 5.393333333333334e-06, "loss": 0.0001, "step": 2575 }, { "epoch": 48.0, "learning_rate": 5.337777777777779e-06, "loss": 0.0001, "step": 2600 }, { "epoch": 48.01, "learning_rate": 5.282222222222223e-06, "loss": 0.0001, "step": 2625 }, { "epoch": 49.0, "learning_rate": 5.226666666666667e-06, "loss": 0.0001, "step": 2650 }, { "epoch": 49.01, "learning_rate": 5.171111111111111e-06, "loss": 0.0001, "step": 2675 }, { "epoch": 49.01, "learning_rate": 5.115555555555556e-06, "loss": 0.0001, "step": 2700 }, { "epoch": 50.01, "learning_rate": 5.060000000000001e-06, "loss": 0.0001, "step": 2725 }, { "epoch": 50.01, "learning_rate": 5.004444444444445e-06, "loss": 0.0001, "step": 2750 }, { "epoch": 51.0, "learning_rate": 4.94888888888889e-06, "loss": 0.0001, "step": 2775 }, { "epoch": 51.01, "learning_rate": 4.893333333333334e-06, "loss": 0.0001, "step": 2800 }, { "epoch": 52.0, "learning_rate": 4.837777777777778e-06, "loss": 0.0001, "step": 2825 }, { "epoch": 52.01, "learning_rate": 4.7822222222222226e-06, "loss": 0.0, "step": 2850 }, { "epoch": 53.0, "learning_rate": 4.7266666666666674e-06, "loss": 0.0001, "step": 2875 }, { "epoch": 53.01, "learning_rate": 4.6711111111111115e-06, "loss": 0.0, "step": 2900 }, { "epoch": 54.0, "learning_rate": 4.6155555555555555e-06, "loss": 0.0, "step": 2925 }, { "epoch": 54.01, "learning_rate": 4.56e-06, "loss": 0.0, "step": 2950 }, { "epoch": 55.0, "learning_rate": 4.504444444444444e-06, "loss": 0.0, "step": 2975 }, { "epoch": 55.01, "learning_rate": 4.448888888888889e-06, "loss": 0.0, "step": 3000 }, { "epoch": 55.01, "eval_loss": 0.6786662936210632, "eval_runtime": 105.9034, "eval_samples_per_second": 3.626, "eval_steps_per_second": 0.227, "eval_wer": 29.816700610997966, "step": 3000 }, { "epoch": 56.0, "learning_rate": 4.393333333333334e-06, "loss": 0.0, "step": 3025 }, { "epoch": 56.01, "learning_rate": 4.337777777777778e-06, "loss": 0.0, "step": 3050 }, { "epoch": 56.01, "learning_rate": 4.282222222222222e-06, "loss": 0.0, "step": 3075 }, { "epoch": 57.0, "learning_rate": 4.226666666666667e-06, "loss": 0.0, "step": 3100 }, { "epoch": 57.01, "learning_rate": 4.171111111111111e-06, "loss": 0.0, "step": 3125 }, { "epoch": 58.0, "learning_rate": 4.115555555555556e-06, "loss": 0.0, "step": 3150 }, { "epoch": 58.01, "learning_rate": 4.060000000000001e-06, "loss": 0.0, "step": 3175 }, { "epoch": 59.0, "learning_rate": 4.004444444444445e-06, "loss": 0.0, "step": 3200 }, { "epoch": 59.01, "learning_rate": 3.948888888888889e-06, "loss": 0.0, "step": 3225 }, { "epoch": 60.0, "learning_rate": 3.893333333333333e-06, "loss": 0.0, "step": 3250 }, { "epoch": 60.01, "learning_rate": 3.837777777777778e-06, "loss": 0.0, "step": 3275 }, { "epoch": 61.0, "learning_rate": 3.782222222222223e-06, "loss": 0.0, "step": 3300 }, { "epoch": 61.01, "learning_rate": 3.726666666666667e-06, "loss": 0.0, "step": 3325 }, { "epoch": 62.0, "learning_rate": 3.6711111111111113e-06, "loss": 0.0, "step": 3350 }, { "epoch": 62.01, "learning_rate": 3.615555555555556e-06, "loss": 0.0, "step": 3375 }, { "epoch": 62.01, "learning_rate": 3.5600000000000002e-06, "loss": 0.0, "step": 3400 }, { "epoch": 63.0, "learning_rate": 3.5044444444444447e-06, "loss": 0.0, "step": 3425 }, { "epoch": 63.01, "learning_rate": 3.4488888888888896e-06, "loss": 0.0, "step": 3450 }, { "epoch": 64.0, "learning_rate": 3.3933333333333336e-06, "loss": 0.0, "step": 3475 }, { "epoch": 64.01, "learning_rate": 3.337777777777778e-06, "loss": 0.0, "step": 3500 }, { "epoch": 64.01, "eval_loss": 0.6893557906150818, "eval_runtime": 106.2501, "eval_samples_per_second": 3.614, "eval_steps_per_second": 0.226, "eval_wer": 29.65376782077393, "step": 3500 }, { "epoch": 65.0, "learning_rate": 3.282222222222223e-06, "loss": 0.0, "step": 3525 }, { "epoch": 65.01, "learning_rate": 3.226666666666667e-06, "loss": 0.0, "step": 3550 }, { "epoch": 66.0, "learning_rate": 3.1711111111111114e-06, "loss": 0.0, "step": 3575 }, { "epoch": 66.01, "learning_rate": 3.1155555555555555e-06, "loss": 0.0, "step": 3600 }, { "epoch": 67.0, "learning_rate": 3.0600000000000003e-06, "loss": 0.0, "step": 3625 }, { "epoch": 67.01, "learning_rate": 3.004444444444445e-06, "loss": 0.0, "step": 3650 }, { "epoch": 68.0, "learning_rate": 2.948888888888889e-06, "loss": 0.0, "step": 3675 }, { "epoch": 68.01, "learning_rate": 2.8933333333333337e-06, "loss": 0.0, "step": 3700 }, { "epoch": 68.01, "learning_rate": 2.837777777777778e-06, "loss": 0.0, "step": 3725 }, { "epoch": 69.0, "learning_rate": 2.7822222222222222e-06, "loss": 0.0, "step": 3750 }, { "epoch": 69.01, "learning_rate": 2.726666666666667e-06, "loss": 0.0, "step": 3775 }, { "epoch": 70.0, "learning_rate": 2.6711111111111116e-06, "loss": 0.0, "step": 3800 }, { "epoch": 70.01, "learning_rate": 2.6155555555555556e-06, "loss": 0.0, "step": 3825 }, { "epoch": 71.0, "learning_rate": 2.56e-06, "loss": 0.0, "step": 3850 }, { "epoch": 71.01, "learning_rate": 2.504444444444445e-06, "loss": 0.0, "step": 3875 }, { "epoch": 72.0, "learning_rate": 2.448888888888889e-06, "loss": 0.0, "step": 3900 }, { "epoch": 72.01, "learning_rate": 2.3933333333333334e-06, "loss": 0.0, "step": 3925 }, { "epoch": 73.0, "learning_rate": 2.337777777777778e-06, "loss": 0.0, "step": 3950 }, { "epoch": 73.01, "learning_rate": 2.2822222222222223e-06, "loss": 0.0, "step": 3975 }, { "epoch": 74.0, "learning_rate": 2.226666666666667e-06, "loss": 0.0, "step": 4000 }, { "epoch": 74.0, "eval_loss": 0.6969531178474426, "eval_runtime": 104.8702, "eval_samples_per_second": 3.662, "eval_steps_per_second": 0.229, "eval_wer": 29.65376782077393, "step": 4000 }, { "epoch": 74.01, "learning_rate": 2.1711111111111113e-06, "loss": 0.0, "step": 4025 }, { "epoch": 74.01, "learning_rate": 2.1155555555555557e-06, "loss": 0.0, "step": 4050 }, { "epoch": 75.0, "learning_rate": 2.06e-06, "loss": 0.0, "step": 4075 }, { "epoch": 75.01, "learning_rate": 2.0044444444444446e-06, "loss": 0.0, "step": 4100 }, { "epoch": 76.0, "learning_rate": 1.948888888888889e-06, "loss": 0.0, "step": 4125 }, { "epoch": 76.01, "learning_rate": 1.8933333333333333e-06, "loss": 0.0, "step": 4150 }, { "epoch": 77.0, "learning_rate": 1.837777777777778e-06, "loss": 0.0, "step": 4175 }, { "epoch": 77.01, "learning_rate": 1.7822222222222225e-06, "loss": 0.0, "step": 4200 }, { "epoch": 78.0, "learning_rate": 1.7266666666666667e-06, "loss": 0.0, "step": 4225 }, { "epoch": 78.01, "learning_rate": 1.6711111111111112e-06, "loss": 0.0, "step": 4250 }, { "epoch": 79.0, "learning_rate": 1.6155555555555559e-06, "loss": 0.0, "step": 4275 }, { "epoch": 79.01, "learning_rate": 1.56e-06, "loss": 0.0, "step": 4300 }, { "epoch": 80.0, "learning_rate": 1.5044444444444446e-06, "loss": 0.0, "step": 4325 }, { "epoch": 80.01, "learning_rate": 1.4488888888888892e-06, "loss": 0.0, "step": 4350 }, { "epoch": 81.0, "learning_rate": 1.3933333333333335e-06, "loss": 0.0, "step": 4375 }, { "epoch": 81.01, "learning_rate": 1.337777777777778e-06, "loss": 0.0, "step": 4400 }, { "epoch": 81.01, "learning_rate": 1.2822222222222222e-06, "loss": 0.0, "step": 4425 }, { "epoch": 82.0, "learning_rate": 1.2266666666666666e-06, "loss": 0.0, "step": 4450 }, { "epoch": 82.01, "learning_rate": 1.171111111111111e-06, "loss": 0.0, "step": 4475 }, { "epoch": 83.0, "learning_rate": 1.1155555555555558e-06, "loss": 0.0, "step": 4500 }, { "epoch": 83.0, "eval_loss": 0.7019144892692566, "eval_runtime": 104.9213, "eval_samples_per_second": 3.66, "eval_steps_per_second": 0.229, "eval_wer": 29.694501018329937, "step": 4500 }, { "epoch": 83.01, "learning_rate": 1.06e-06, "loss": 0.0, "step": 4525 }, { "epoch": 84.0, "learning_rate": 1.0044444444444445e-06, "loss": 0.0, "step": 4550 }, { "epoch": 84.01, "learning_rate": 9.488888888888889e-07, "loss": 0.0, "step": 4575 }, { "epoch": 85.0, "learning_rate": 8.933333333333334e-07, "loss": 0.0, "step": 4600 }, { "epoch": 85.01, "learning_rate": 8.37777777777778e-07, "loss": 0.0, "step": 4625 }, { "epoch": 86.0, "learning_rate": 7.822222222222223e-07, "loss": 0.0, "step": 4650 }, { "epoch": 86.01, "learning_rate": 7.266666666666668e-07, "loss": 0.0, "step": 4675 }, { "epoch": 87.0, "learning_rate": 6.711111111111111e-07, "loss": 0.0, "step": 4700 }, { "epoch": 87.01, "learning_rate": 6.155555555555556e-07, "loss": 0.0, "step": 4725 }, { "epoch": 87.01, "learning_rate": 5.6e-07, "loss": 0.0, "step": 4750 }, { "epoch": 88.0, "learning_rate": 5.044444444444445e-07, "loss": 0.0, "step": 4775 }, { "epoch": 88.01, "learning_rate": 4.488888888888889e-07, "loss": 0.0, "step": 4800 }, { "epoch": 89.0, "learning_rate": 3.9333333333333336e-07, "loss": 0.0, "step": 4825 }, { "epoch": 89.01, "learning_rate": 3.3777777777777777e-07, "loss": 0.0, "step": 4850 }, { "epoch": 90.0, "learning_rate": 2.822222222222222e-07, "loss": 0.0, "step": 4875 }, { "epoch": 90.01, "learning_rate": 2.266666666666667e-07, "loss": 0.0, "step": 4900 }, { "epoch": 91.0, "learning_rate": 1.7111111111111114e-07, "loss": 0.0, "step": 4925 }, { "epoch": 91.01, "learning_rate": 1.1555555555555556e-07, "loss": 0.0, "step": 4950 }, { "epoch": 92.0, "learning_rate": 6.000000000000001e-08, "loss": 0.0, "step": 4975 }, { "epoch": 92.01, "learning_rate": 4.444444444444445e-09, "loss": 0.0, "step": 5000 }, { "epoch": 92.01, "eval_loss": 0.7038618922233582, "eval_runtime": 106.4448, "eval_samples_per_second": 3.608, "eval_steps_per_second": 0.225, "eval_wer": 29.735234215885946, "step": 5000 }, { "epoch": 92.01, "step": 5000, "total_flos": 2.717998645248e+20, "train_loss": 0.028814784878876527, "train_runtime": 28148.7573, "train_samples_per_second": 2.842, "train_steps_per_second": 0.178 } ], "logging_steps": 25, "max_steps": 5000, "num_input_tokens_seen": 0, "num_train_epochs": 9223372036854775807, "save_steps": 1000, "total_flos": 2.717998645248e+20, "train_batch_size": 16, "trial_name": null, "trial_params": null }