{ "best_metric": 17.189821693907874, "best_model_checkpoint": "./checkpoint-5000", "epoch": 46.728971962616825, "global_step": 5000, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.23, "learning_rate": 1.5136083400296205e-06, "loss": 0.0026, "step": 25 }, { "epoch": 0.47, "learning_rate": 1.8687587131475301e-06, "loss": 0.0022, "step": 50 }, { "epoch": 0.7, "learning_rate": 2.0711488350670174e-06, "loss": 0.0021, "step": 75 }, { "epoch": 0.93, "learning_rate": 2.213317753617305e-06, "loss": 0.0021, "step": 100 }, { "epoch": 1.17, "learning_rate": 2.3230029693718747e-06, "loss": 0.0016, "step": 125 }, { "epoch": 1.4, "learning_rate": 2.412322158351148e-06, "loss": 0.0017, "step": 150 }, { "epoch": 1.64, "learning_rate": 2.4876668872198717e-06, "loss": 0.0017, "step": 175 }, { "epoch": 1.87, "learning_rate": 2.552824062407326e-06, "loss": 0.0014, "step": 200 }, { "epoch": 2.1, "learning_rate": 2.610223373296667e-06, "loss": 0.0015, "step": 225 }, { "epoch": 2.34, "learning_rate": 2.661517182828361e-06, "loss": 0.0017, "step": 250 }, { "epoch": 2.57, "learning_rate": 2.7078803874740543e-06, "loss": 0.002, "step": 275 }, { "epoch": 2.8, "learning_rate": 2.750178319990197e-06, "loss": 0.0021, "step": 300 }, { "epoch": 3.04, "learning_rate": 2.7890667754365044e-06, "loss": 0.0016, "step": 325 }, { "epoch": 3.27, "learning_rate": 2.8250546392106077e-06, "loss": 0.0013, "step": 350 }, { "epoch": 3.5, "learning_rate": 2.8585447348549113e-06, "loss": 0.0017, "step": 375 }, { "epoch": 3.74, "learning_rate": 2.889861392935294e-06, "loss": 0.0014, "step": 400 }, { "epoch": 3.97, "learning_rate": 2.9192696063561725e-06, "loss": 0.0015, "step": 425 }, { "epoch": 4.21, "learning_rate": 2.946988676871634e-06, "loss": 0.0013, "step": 450 }, { "epoch": 4.44, "learning_rate": 2.973202150939645e-06, "loss": 0.0014, "step": 475 }, { "epoch": 4.67, "learning_rate": 2.998065193492142e-06, "loss": 0.0013, "step": 500 }, { "epoch": 4.91, "learning_rate": 3e-06, "loss": 0.0019, "step": 525 }, { "epoch": 5.14, "learning_rate": 3e-06, "loss": 0.0018, "step": 550 }, { "epoch": 5.37, "learning_rate": 3e-06, "loss": 0.0016, "step": 575 }, { "epoch": 5.61, "learning_rate": 3e-06, "loss": 0.002, "step": 600 }, { "epoch": 5.84, "learning_rate": 3e-06, "loss": 0.0023, "step": 625 }, { "epoch": 6.07, "learning_rate": 3e-06, "loss": 0.0014, "step": 650 }, { "epoch": 6.31, "learning_rate": 3e-06, "loss": 0.0013, "step": 675 }, { "epoch": 6.54, "learning_rate": 3e-06, "loss": 0.0013, "step": 700 }, { "epoch": 6.78, "learning_rate": 3e-06, "loss": 0.0013, "step": 725 }, { "epoch": 7.01, "learning_rate": 3e-06, "loss": 0.0019, "step": 750 }, { "epoch": 7.24, "learning_rate": 3e-06, "loss": 0.0017, "step": 775 }, { "epoch": 7.48, "learning_rate": 3e-06, "loss": 0.0016, "step": 800 }, { "epoch": 7.71, "learning_rate": 3e-06, "loss": 0.0022, "step": 825 }, { "epoch": 7.94, "learning_rate": 3e-06, "loss": 0.0017, "step": 850 }, { "epoch": 8.18, "learning_rate": 3e-06, "loss": 0.0014, "step": 875 }, { "epoch": 8.41, "learning_rate": 3e-06, "loss": 0.0015, "step": 900 }, { "epoch": 8.64, "learning_rate": 3e-06, "loss": 0.0017, "step": 925 }, { "epoch": 8.88, "learning_rate": 3e-06, "loss": 0.0016, "step": 950 }, { "epoch": 9.11, "learning_rate": 3e-06, "loss": 0.0011, "step": 975 }, { "epoch": 9.35, "learning_rate": 3e-06, "loss": 0.0015, "step": 1000 }, { "epoch": 9.35, "eval_loss": 0.54052734375, "eval_runtime": 168.3935, "eval_samples_per_second": 10.072, "eval_steps_per_second": 0.629, "eval_wer": 17.691307578008917, "step": 1000 }, { "epoch": 9.58, "learning_rate": 3e-06, "loss": 0.0027, "step": 1025 }, { "epoch": 9.81, "learning_rate": 3e-06, "loss": 0.0013, "step": 1050 }, { "epoch": 10.05, "learning_rate": 3e-06, "loss": 0.0015, "step": 1075 }, { "epoch": 10.28, "learning_rate": 3e-06, "loss": 0.0016, "step": 1100 }, { "epoch": 10.51, "learning_rate": 3e-06, "loss": 0.0011, "step": 1125 }, { "epoch": 10.75, "learning_rate": 3e-06, "loss": 0.0017, "step": 1150 }, { "epoch": 10.98, "learning_rate": 3e-06, "loss": 0.0018, "step": 1175 }, { "epoch": 11.21, "learning_rate": 3e-06, "loss": 0.0017, "step": 1200 }, { "epoch": 11.45, "learning_rate": 3e-06, "loss": 0.0021, "step": 1225 }, { "epoch": 11.68, "learning_rate": 3e-06, "loss": 0.0017, "step": 1250 }, { "epoch": 11.92, "learning_rate": 3e-06, "loss": 0.0015, "step": 1275 }, { "epoch": 12.15, "learning_rate": 3e-06, "loss": 0.0012, "step": 1300 }, { "epoch": 12.38, "learning_rate": 3e-06, "loss": 0.0016, "step": 1325 }, { "epoch": 12.62, "learning_rate": 3e-06, "loss": 0.001, "step": 1350 }, { "epoch": 12.85, "learning_rate": 3e-06, "loss": 0.0013, "step": 1375 }, { "epoch": 13.08, "learning_rate": 3e-06, "loss": 0.0013, "step": 1400 }, { "epoch": 13.32, "learning_rate": 3e-06, "loss": 0.0011, "step": 1425 }, { "epoch": 13.55, "learning_rate": 3e-06, "loss": 0.0013, "step": 1450 }, { "epoch": 13.79, "learning_rate": 3e-06, "loss": 0.0015, "step": 1475 }, { "epoch": 14.02, "learning_rate": 3e-06, "loss": 0.0016, "step": 1500 }, { "epoch": 14.25, "learning_rate": 3e-06, "loss": 0.0013, "step": 1525 }, { "epoch": 14.49, "learning_rate": 3e-06, "loss": 0.0011, "step": 1550 }, { "epoch": 14.72, "learning_rate": 3e-06, "loss": 0.0014, "step": 1575 }, { "epoch": 14.95, "learning_rate": 3e-06, "loss": 0.0016, "step": 1600 }, { "epoch": 15.19, "learning_rate": 3e-06, "loss": 0.0013, "step": 1625 }, { "epoch": 15.42, "learning_rate": 3e-06, "loss": 0.0013, "step": 1650 }, { "epoch": 15.65, "learning_rate": 3e-06, "loss": 0.0011, "step": 1675 }, { "epoch": 15.89, "learning_rate": 3e-06, "loss": 0.0012, "step": 1700 }, { "epoch": 16.12, "learning_rate": 3e-06, "loss": 0.0013, "step": 1725 }, { "epoch": 16.36, "learning_rate": 3e-06, "loss": 0.0012, "step": 1750 }, { "epoch": 16.59, "learning_rate": 3e-06, "loss": 0.001, "step": 1775 }, { "epoch": 16.82, "learning_rate": 3e-06, "loss": 0.0014, "step": 1800 }, { "epoch": 17.06, "learning_rate": 3e-06, "loss": 0.0016, "step": 1825 }, { "epoch": 17.29, "learning_rate": 3e-06, "loss": 0.0011, "step": 1850 }, { "epoch": 17.52, "learning_rate": 3e-06, "loss": 0.0015, "step": 1875 }, { "epoch": 17.76, "learning_rate": 3e-06, "loss": 0.0014, "step": 1900 }, { "epoch": 17.99, "learning_rate": 3e-06, "loss": 0.0012, "step": 1925 }, { "epoch": 18.22, "learning_rate": 3e-06, "loss": 0.001, "step": 1950 }, { "epoch": 18.46, "learning_rate": 3e-06, "loss": 0.0009, "step": 1975 }, { "epoch": 18.69, "learning_rate": 3e-06, "loss": 0.001, "step": 2000 }, { "epoch": 18.69, "eval_loss": 0.53955078125, "eval_runtime": 168.0749, "eval_samples_per_second": 10.091, "eval_steps_per_second": 0.631, "eval_wer": 17.561292719167902, "step": 2000 }, { "epoch": 18.93, "learning_rate": 3e-06, "loss": 0.0015, "step": 2025 }, { "epoch": 19.16, "learning_rate": 3e-06, "loss": 0.0011, "step": 2050 }, { "epoch": 19.39, "learning_rate": 3e-06, "loss": 0.0013, "step": 2075 }, { "epoch": 19.63, "learning_rate": 3e-06, "loss": 0.0012, "step": 2100 }, { "epoch": 19.86, "learning_rate": 3e-06, "loss": 0.0017, "step": 2125 }, { "epoch": 20.09, "learning_rate": 3e-06, "loss": 0.0009, "step": 2150 }, { "epoch": 20.33, "learning_rate": 3e-06, "loss": 0.0009, "step": 2175 }, { "epoch": 20.56, "learning_rate": 3e-06, "loss": 0.0014, "step": 2200 }, { "epoch": 20.79, "learning_rate": 3e-06, "loss": 0.0014, "step": 2225 }, { "epoch": 21.03, "learning_rate": 3e-06, "loss": 0.0009, "step": 2250 }, { "epoch": 21.26, "learning_rate": 3e-06, "loss": 0.0011, "step": 2275 }, { "epoch": 21.5, "learning_rate": 3e-06, "loss": 0.0016, "step": 2300 }, { "epoch": 21.73, "learning_rate": 3e-06, "loss": 0.0017, "step": 2325 }, { "epoch": 21.96, "learning_rate": 3e-06, "loss": 0.0012, "step": 2350 }, { "epoch": 22.2, "learning_rate": 3e-06, "loss": 0.0013, "step": 2375 }, { "epoch": 22.43, "learning_rate": 3e-06, "loss": 0.0013, "step": 2400 }, { "epoch": 22.66, "learning_rate": 3e-06, "loss": 0.0013, "step": 2425 }, { "epoch": 22.9, "learning_rate": 3e-06, "loss": 0.0015, "step": 2450 }, { "epoch": 23.13, "learning_rate": 3e-06, "loss": 0.0015, "step": 2475 }, { "epoch": 23.36, "learning_rate": 3e-06, "loss": 0.001, "step": 2500 }, { "epoch": 23.6, "learning_rate": 3e-06, "loss": 0.0009, "step": 2525 }, { "epoch": 23.83, "learning_rate": 3e-06, "loss": 0.0009, "step": 2550 }, { "epoch": 24.07, "learning_rate": 3e-06, "loss": 0.0008, "step": 2575 }, { "epoch": 24.3, "learning_rate": 3e-06, "loss": 0.0008, "step": 2600 }, { "epoch": 24.53, "learning_rate": 3e-06, "loss": 0.0012, "step": 2625 }, { "epoch": 24.77, "learning_rate": 3e-06, "loss": 0.0008, "step": 2650 }, { "epoch": 25.0, "learning_rate": 3e-06, "loss": 0.0008, "step": 2675 }, { "epoch": 25.23, "learning_rate": 3e-06, "loss": 0.001, "step": 2700 }, { "epoch": 25.47, "learning_rate": 3e-06, "loss": 0.0008, "step": 2725 }, { "epoch": 25.7, "learning_rate": 3e-06, "loss": 0.0009, "step": 2750 }, { "epoch": 25.93, "learning_rate": 3e-06, "loss": 0.0008, "step": 2775 }, { "epoch": 26.17, "learning_rate": 3e-06, "loss": 0.0012, "step": 2800 }, { "epoch": 26.4, "learning_rate": 3e-06, "loss": 0.0007, "step": 2825 }, { "epoch": 26.64, "learning_rate": 3e-06, "loss": 0.001, "step": 2850 }, { "epoch": 26.87, "learning_rate": 3e-06, "loss": 0.0009, "step": 2875 }, { "epoch": 27.1, "learning_rate": 3e-06, "loss": 0.0011, "step": 2900 }, { "epoch": 27.34, "learning_rate": 3e-06, "loss": 0.001, "step": 2925 }, { "epoch": 27.57, "learning_rate": 3e-06, "loss": 0.0009, "step": 2950 }, { "epoch": 27.8, "learning_rate": 3e-06, "loss": 0.001, "step": 2975 }, { "epoch": 28.04, "learning_rate": 3e-06, "loss": 0.0009, "step": 3000 }, { "epoch": 28.04, "eval_loss": 0.56201171875, "eval_runtime": 168.2778, "eval_samples_per_second": 10.079, "eval_steps_per_second": 0.63, "eval_wer": 17.496285289747398, "step": 3000 }, { "epoch": 28.27, "learning_rate": 3e-06, "loss": 0.0015, "step": 3025 }, { "epoch": 28.5, "learning_rate": 3e-06, "loss": 0.001, "step": 3050 }, { "epoch": 28.74, "learning_rate": 3e-06, "loss": 0.0008, "step": 3075 }, { "epoch": 28.97, "learning_rate": 3e-06, "loss": 0.0011, "step": 3100 }, { "epoch": 29.21, "learning_rate": 3e-06, "loss": 0.0009, "step": 3125 }, { "epoch": 29.44, "learning_rate": 3e-06, "loss": 0.0013, "step": 3150 }, { "epoch": 29.67, "learning_rate": 3e-06, "loss": 0.0009, "step": 3175 }, { "epoch": 29.91, "learning_rate": 3e-06, "loss": 0.0011, "step": 3200 }, { "epoch": 30.14, "learning_rate": 3e-06, "loss": 0.0013, "step": 3225 }, { "epoch": 30.37, "learning_rate": 3e-06, "loss": 0.0012, "step": 3250 }, { "epoch": 30.61, "learning_rate": 3e-06, "loss": 0.0012, "step": 3275 }, { "epoch": 30.84, "learning_rate": 3e-06, "loss": 0.0008, "step": 3300 }, { "epoch": 31.07, "learning_rate": 3e-06, "loss": 0.0013, "step": 3325 }, { "epoch": 31.31, "learning_rate": 3e-06, "loss": 0.0009, "step": 3350 }, { "epoch": 31.54, "learning_rate": 3e-06, "loss": 0.0008, "step": 3375 }, { "epoch": 31.78, "learning_rate": 3e-06, "loss": 0.0009, "step": 3400 }, { "epoch": 32.01, "learning_rate": 3e-06, "loss": 0.002, "step": 3425 }, { "epoch": 32.24, "learning_rate": 3e-06, "loss": 0.0009, "step": 3450 }, { "epoch": 32.48, "learning_rate": 3e-06, "loss": 0.0014, "step": 3475 }, { "epoch": 32.71, "learning_rate": 3e-06, "loss": 0.0009, "step": 3500 }, { "epoch": 32.94, "learning_rate": 3e-06, "loss": 0.0013, "step": 3525 }, { "epoch": 33.18, "learning_rate": 3e-06, "loss": 0.001, "step": 3550 }, { "epoch": 33.41, "learning_rate": 3e-06, "loss": 0.0008, "step": 3575 }, { "epoch": 33.64, "learning_rate": 3e-06, "loss": 0.0012, "step": 3600 }, { "epoch": 33.88, "learning_rate": 3e-06, "loss": 0.0011, "step": 3625 }, { "epoch": 34.11, "learning_rate": 3e-06, "loss": 0.0012, "step": 3650 }, { "epoch": 34.35, "learning_rate": 3e-06, "loss": 0.0008, "step": 3675 }, { "epoch": 34.58, "learning_rate": 3e-06, "loss": 0.0017, "step": 3700 }, { "epoch": 34.81, "learning_rate": 3e-06, "loss": 0.0018, "step": 3725 }, { "epoch": 35.05, "learning_rate": 3e-06, "loss": 0.0008, "step": 3750 }, { "epoch": 35.28, "learning_rate": 3e-06, "loss": 0.0012, "step": 3775 }, { "epoch": 35.51, "learning_rate": 3e-06, "loss": 0.0026, "step": 3800 }, { "epoch": 35.75, "learning_rate": 3e-06, "loss": 0.0013, "step": 3825 }, { "epoch": 35.98, "learning_rate": 3e-06, "loss": 0.0015, "step": 3850 }, { "epoch": 36.21, "learning_rate": 3e-06, "loss": 0.0012, "step": 3875 }, { "epoch": 36.45, "learning_rate": 3e-06, "loss": 0.0017, "step": 3900 }, { "epoch": 36.68, "learning_rate": 3e-06, "loss": 0.0009, "step": 3925 }, { "epoch": 36.92, "learning_rate": 3e-06, "loss": 0.0017, "step": 3950 }, { "epoch": 37.15, "learning_rate": 3e-06, "loss": 0.0007, "step": 3975 }, { "epoch": 37.38, "learning_rate": 3e-06, "loss": 0.001, "step": 4000 }, { "epoch": 37.38, "eval_loss": 0.54541015625, "eval_runtime": 168.5354, "eval_samples_per_second": 10.063, "eval_steps_per_second": 0.629, "eval_wer": 17.459138187221395, "step": 4000 }, { "epoch": 37.62, "learning_rate": 3e-06, "loss": 0.0013, "step": 4025 }, { "epoch": 37.85, "learning_rate": 3e-06, "loss": 0.0016, "step": 4050 }, { "epoch": 38.08, "learning_rate": 3e-06, "loss": 0.0018, "step": 4075 }, { "epoch": 38.32, "learning_rate": 3e-06, "loss": 0.0017, "step": 4100 }, { "epoch": 38.55, "learning_rate": 3e-06, "loss": 0.0023, "step": 4125 }, { "epoch": 38.79, "learning_rate": 3e-06, "loss": 0.0019, "step": 4150 }, { "epoch": 39.02, "learning_rate": 3e-06, "loss": 0.0026, "step": 4175 }, { "epoch": 39.25, "learning_rate": 3e-06, "loss": 0.0025, "step": 4200 }, { "epoch": 39.49, "learning_rate": 3e-06, "loss": 0.0022, "step": 4225 }, { "epoch": 39.72, "learning_rate": 3e-06, "loss": 0.0018, "step": 4250 }, { "epoch": 39.95, "learning_rate": 3e-06, "loss": 0.0024, "step": 4275 }, { "epoch": 40.19, "learning_rate": 3e-06, "loss": 0.0023, "step": 4300 }, { "epoch": 40.42, "learning_rate": 3e-06, "loss": 0.0017, "step": 4325 }, { "epoch": 40.65, "learning_rate": 3e-06, "loss": 0.0023, "step": 4350 }, { "epoch": 40.89, "learning_rate": 3e-06, "loss": 0.0029, "step": 4375 }, { "epoch": 41.12, "learning_rate": 3e-06, "loss": 0.0024, "step": 4400 }, { "epoch": 41.36, "learning_rate": 3e-06, "loss": 0.0016, "step": 4425 }, { "epoch": 41.59, "learning_rate": 3e-06, "loss": 0.0022, "step": 4450 }, { "epoch": 41.82, "learning_rate": 3e-06, "loss": 0.0021, "step": 4475 }, { "epoch": 42.06, "learning_rate": 3e-06, "loss": 0.0026, "step": 4500 }, { "epoch": 42.29, "learning_rate": 3e-06, "loss": 0.0023, "step": 4525 }, { "epoch": 42.52, "learning_rate": 3e-06, "loss": 0.0019, "step": 4550 }, { "epoch": 42.76, "learning_rate": 3e-06, "loss": 0.0025, "step": 4575 }, { "epoch": 42.99, "learning_rate": 3e-06, "loss": 0.0019, "step": 4600 }, { "epoch": 43.22, "learning_rate": 3e-06, "loss": 0.0023, "step": 4625 }, { "epoch": 43.46, "learning_rate": 3e-06, "loss": 0.0022, "step": 4650 }, { "epoch": 43.69, "learning_rate": 3e-06, "loss": 0.0032, "step": 4675 }, { "epoch": 43.93, "learning_rate": 3e-06, "loss": 0.0021, "step": 4700 }, { "epoch": 44.16, "learning_rate": 3e-06, "loss": 0.0017, "step": 4725 }, { "epoch": 44.39, "learning_rate": 3e-06, "loss": 0.0022, "step": 4750 }, { "epoch": 44.63, "learning_rate": 3e-06, "loss": 0.0026, "step": 4775 }, { "epoch": 44.86, "learning_rate": 3e-06, "loss": 0.002, "step": 4800 }, { "epoch": 45.09, "learning_rate": 3e-06, "loss": 0.0019, "step": 4825 }, { "epoch": 45.33, "learning_rate": 3e-06, "loss": 0.0022, "step": 4850 }, { "epoch": 45.56, "learning_rate": 3e-06, "loss": 0.0016, "step": 4875 }, { "epoch": 45.79, "learning_rate": 3e-06, "loss": 0.0029, "step": 4900 }, { "epoch": 46.03, "learning_rate": 3e-06, "loss": 0.0019, "step": 4925 }, { "epoch": 46.26, "learning_rate": 3e-06, "loss": 0.0023, "step": 4950 }, { "epoch": 46.5, "learning_rate": 3e-06, "loss": 0.0019, "step": 4975 }, { "epoch": 46.73, "learning_rate": 3e-06, "loss": 0.0015, "step": 5000 }, { "epoch": 46.73, "eval_loss": 0.54296875, "eval_runtime": 168.1004, "eval_samples_per_second": 10.089, "eval_steps_per_second": 0.631, "eval_wer": 17.189821693907874, "step": 5000 }, { "epoch": 46.73, "step": 5000, "total_flos": 4.5881616269897105e+19, "train_loss": 0.0014968906164169312, "train_runtime": 9420.014, "train_samples_per_second": 16.985, "train_steps_per_second": 0.531 } ], "max_steps": 5000, "num_train_epochs": 47, "total_flos": 4.5881616269897105e+19, "trial_name": null, "trial_params": null }