| { | |
| "best_metric": 18.056954491346946, | |
| "best_model_checkpoint": "./checkpoint-4000", | |
| "epoch": 48.899755501222494, | |
| "eval_steps": 1000, | |
| "global_step": 5000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.24, | |
| "grad_norm": 39.86528778076172, | |
| "learning_rate": 5.000000000000001e-07, | |
| "loss": 2.0555, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "grad_norm": Infinity, | |
| "learning_rate": 9.800000000000001e-07, | |
| "loss": 1.5219, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "grad_norm": 6.200109958648682, | |
| "learning_rate": 1.48e-06, | |
| "loss": 1.0167, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "grad_norm": 5.486103057861328, | |
| "learning_rate": 1.98e-06, | |
| "loss": 0.7299, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.22, | |
| "grad_norm": 5.532894134521484, | |
| "learning_rate": 2.4800000000000004e-06, | |
| "loss": 0.6318, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 1.47, | |
| "grad_norm": 4.895535945892334, | |
| "learning_rate": 2.9800000000000003e-06, | |
| "loss": 0.5503, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.71, | |
| "grad_norm": 5.065937519073486, | |
| "learning_rate": 3.48e-06, | |
| "loss": 0.4999, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 1.96, | |
| "grad_norm": 4.807600498199463, | |
| "learning_rate": 3.980000000000001e-06, | |
| "loss": 0.4457, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 2.2, | |
| "grad_norm": 4.568106174468994, | |
| "learning_rate": 4.48e-06, | |
| "loss": 0.3779, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 2.44, | |
| "grad_norm": 4.705833911895752, | |
| "learning_rate": 4.980000000000001e-06, | |
| "loss": 0.3332, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 2.69, | |
| "grad_norm": 4.315070629119873, | |
| "learning_rate": 5.480000000000001e-06, | |
| "loss": 0.2858, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 2.93, | |
| "grad_norm": 2.660728693008423, | |
| "learning_rate": 5.98e-06, | |
| "loss": 0.2228, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 3.18, | |
| "grad_norm": 2.5569632053375244, | |
| "learning_rate": 6.480000000000001e-06, | |
| "loss": 0.1766, | |
| "step": 325 | |
| }, | |
| { | |
| "epoch": 3.42, | |
| "grad_norm": 2.2663660049438477, | |
| "learning_rate": 6.98e-06, | |
| "loss": 0.1475, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 3.67, | |
| "grad_norm": 2.5257091522216797, | |
| "learning_rate": 7.48e-06, | |
| "loss": 0.1492, | |
| "step": 375 | |
| }, | |
| { | |
| "epoch": 3.91, | |
| "grad_norm": 2.3954405784606934, | |
| "learning_rate": 7.980000000000002e-06, | |
| "loss": 0.142, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 4.16, | |
| "grad_norm": 2.181328296661377, | |
| "learning_rate": 8.48e-06, | |
| "loss": 0.1125, | |
| "step": 425 | |
| }, | |
| { | |
| "epoch": 4.4, | |
| "grad_norm": 1.9877594709396362, | |
| "learning_rate": 8.98e-06, | |
| "loss": 0.0908, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 4.65, | |
| "grad_norm": 1.8853321075439453, | |
| "learning_rate": 9.48e-06, | |
| "loss": 0.0892, | |
| "step": 475 | |
| }, | |
| { | |
| "epoch": 4.89, | |
| "grad_norm": 2.3802549839019775, | |
| "learning_rate": 9.980000000000001e-06, | |
| "loss": 0.091, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 5.13, | |
| "grad_norm": 1.3576879501342773, | |
| "learning_rate": 9.946666666666667e-06, | |
| "loss": 0.0713, | |
| "step": 525 | |
| }, | |
| { | |
| "epoch": 5.38, | |
| "grad_norm": 2.4532103538513184, | |
| "learning_rate": 9.891111111111113e-06, | |
| "loss": 0.0534, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 5.62, | |
| "grad_norm": 1.3736106157302856, | |
| "learning_rate": 9.835555555555556e-06, | |
| "loss": 0.0512, | |
| "step": 575 | |
| }, | |
| { | |
| "epoch": 5.87, | |
| "grad_norm": 2.0095458030700684, | |
| "learning_rate": 9.780000000000001e-06, | |
| "loss": 0.0571, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 6.11, | |
| "grad_norm": 1.2924531698226929, | |
| "learning_rate": 9.724444444444445e-06, | |
| "loss": 0.0453, | |
| "step": 625 | |
| }, | |
| { | |
| "epoch": 6.36, | |
| "grad_norm": 1.0012321472167969, | |
| "learning_rate": 9.66888888888889e-06, | |
| "loss": 0.0292, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 6.6, | |
| "grad_norm": 1.4161434173583984, | |
| "learning_rate": 9.613333333333335e-06, | |
| "loss": 0.0325, | |
| "step": 675 | |
| }, | |
| { | |
| "epoch": 6.85, | |
| "grad_norm": 5.784367084503174, | |
| "learning_rate": 9.557777777777777e-06, | |
| "loss": 0.031, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 7.09, | |
| "grad_norm": 0.8382102251052856, | |
| "learning_rate": 9.502222222222223e-06, | |
| "loss": 0.0247, | |
| "step": 725 | |
| }, | |
| { | |
| "epoch": 7.33, | |
| "grad_norm": 1.2963491678237915, | |
| "learning_rate": 9.446666666666667e-06, | |
| "loss": 0.0162, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 7.58, | |
| "grad_norm": 1.7834402322769165, | |
| "learning_rate": 9.391111111111111e-06, | |
| "loss": 0.0175, | |
| "step": 775 | |
| }, | |
| { | |
| "epoch": 7.82, | |
| "grad_norm": 0.9083292484283447, | |
| "learning_rate": 9.335555555555557e-06, | |
| "loss": 0.0193, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 8.07, | |
| "grad_norm": 0.5552634596824646, | |
| "learning_rate": 9.280000000000001e-06, | |
| "loss": 0.0157, | |
| "step": 825 | |
| }, | |
| { | |
| "epoch": 8.31, | |
| "grad_norm": 1.1231069564819336, | |
| "learning_rate": 9.224444444444445e-06, | |
| "loss": 0.0105, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 8.56, | |
| "grad_norm": 1.206103801727295, | |
| "learning_rate": 9.168888888888889e-06, | |
| "loss": 0.0109, | |
| "step": 875 | |
| }, | |
| { | |
| "epoch": 8.8, | |
| "grad_norm": 0.8872191309928894, | |
| "learning_rate": 9.113333333333335e-06, | |
| "loss": 0.0126, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 9.05, | |
| "grad_norm": 0.7421383261680603, | |
| "learning_rate": 9.057777777777779e-06, | |
| "loss": 0.0107, | |
| "step": 925 | |
| }, | |
| { | |
| "epoch": 9.29, | |
| "grad_norm": 0.7581607103347778, | |
| "learning_rate": 9.002222222222223e-06, | |
| "loss": 0.006, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 9.54, | |
| "grad_norm": 0.6848894953727722, | |
| "learning_rate": 8.946666666666669e-06, | |
| "loss": 0.006, | |
| "step": 975 | |
| }, | |
| { | |
| "epoch": 9.78, | |
| "grad_norm": 1.044122576713562, | |
| "learning_rate": 8.891111111111111e-06, | |
| "loss": 0.0067, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 9.78, | |
| "eval_loss": 0.41375118494033813, | |
| "eval_runtime": 1461.809, | |
| "eval_samples_per_second": 1.98, | |
| "eval_steps_per_second": 0.495, | |
| "eval_wer": 18.938284039923083, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 10.02, | |
| "grad_norm": 0.6757261753082275, | |
| "learning_rate": 8.835555555555557e-06, | |
| "loss": 0.0058, | |
| "step": 1025 | |
| }, | |
| { | |
| "epoch": 10.27, | |
| "grad_norm": 1.085519552230835, | |
| "learning_rate": 8.78e-06, | |
| "loss": 0.0037, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 10.51, | |
| "grad_norm": 0.8559943437576294, | |
| "learning_rate": 8.724444444444445e-06, | |
| "loss": 0.0044, | |
| "step": 1075 | |
| }, | |
| { | |
| "epoch": 10.76, | |
| "grad_norm": 1.7756787538528442, | |
| "learning_rate": 8.66888888888889e-06, | |
| "loss": 0.0056, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "grad_norm": 0.5664415955543518, | |
| "learning_rate": 8.613333333333333e-06, | |
| "loss": 0.0048, | |
| "step": 1125 | |
| }, | |
| { | |
| "epoch": 11.25, | |
| "grad_norm": 0.621498703956604, | |
| "learning_rate": 8.557777777777778e-06, | |
| "loss": 0.0038, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 11.49, | |
| "grad_norm": 0.9859088659286499, | |
| "learning_rate": 8.502222222222223e-06, | |
| "loss": 0.0035, | |
| "step": 1175 | |
| }, | |
| { | |
| "epoch": 11.74, | |
| "grad_norm": 1.2961162328720093, | |
| "learning_rate": 8.446666666666668e-06, | |
| "loss": 0.0041, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 11.98, | |
| "grad_norm": 0.5769420862197876, | |
| "learning_rate": 8.391111111111112e-06, | |
| "loss": 0.0035, | |
| "step": 1225 | |
| }, | |
| { | |
| "epoch": 12.22, | |
| "grad_norm": 0.5504060387611389, | |
| "learning_rate": 8.335555555555556e-06, | |
| "loss": 0.0022, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 12.47, | |
| "grad_norm": 0.7063620090484619, | |
| "learning_rate": 8.28e-06, | |
| "loss": 0.0027, | |
| "step": 1275 | |
| }, | |
| { | |
| "epoch": 12.71, | |
| "grad_norm": 0.6650658845901489, | |
| "learning_rate": 8.224444444444444e-06, | |
| "loss": 0.0029, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 12.96, | |
| "grad_norm": 0.6803381443023682, | |
| "learning_rate": 8.16888888888889e-06, | |
| "loss": 0.0023, | |
| "step": 1325 | |
| }, | |
| { | |
| "epoch": 13.2, | |
| "grad_norm": 0.19391104578971863, | |
| "learning_rate": 8.113333333333334e-06, | |
| "loss": 0.0013, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 13.45, | |
| "grad_norm": 0.43767812848091125, | |
| "learning_rate": 8.057777777777778e-06, | |
| "loss": 0.002, | |
| "step": 1375 | |
| }, | |
| { | |
| "epoch": 13.69, | |
| "grad_norm": 0.6082565188407898, | |
| "learning_rate": 8.002222222222222e-06, | |
| "loss": 0.0022, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 13.94, | |
| "grad_norm": 0.30705004930496216, | |
| "learning_rate": 7.946666666666666e-06, | |
| "loss": 0.002, | |
| "step": 1425 | |
| }, | |
| { | |
| "epoch": 14.18, | |
| "grad_norm": 0.18880507349967957, | |
| "learning_rate": 7.891111111111112e-06, | |
| "loss": 0.0015, | |
| "step": 1450 | |
| }, | |
| { | |
| "epoch": 14.43, | |
| "grad_norm": 0.32524725794792175, | |
| "learning_rate": 7.835555555555556e-06, | |
| "loss": 0.0015, | |
| "step": 1475 | |
| }, | |
| { | |
| "epoch": 14.67, | |
| "grad_norm": 2.48786997795105, | |
| "learning_rate": 7.78e-06, | |
| "loss": 0.0015, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 14.91, | |
| "grad_norm": 0.3373986482620239, | |
| "learning_rate": 7.724444444444446e-06, | |
| "loss": 0.0013, | |
| "step": 1525 | |
| }, | |
| { | |
| "epoch": 15.16, | |
| "grad_norm": 0.29098883271217346, | |
| "learning_rate": 7.66888888888889e-06, | |
| "loss": 0.0011, | |
| "step": 1550 | |
| }, | |
| { | |
| "epoch": 15.4, | |
| "grad_norm": 0.12477891892194748, | |
| "learning_rate": 7.613333333333334e-06, | |
| "loss": 0.0008, | |
| "step": 1575 | |
| }, | |
| { | |
| "epoch": 15.65, | |
| "grad_norm": 0.06489470601081848, | |
| "learning_rate": 7.557777777777779e-06, | |
| "loss": 0.0011, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 15.89, | |
| "grad_norm": 0.061178650707006454, | |
| "learning_rate": 7.502222222222223e-06, | |
| "loss": 0.001, | |
| "step": 1625 | |
| }, | |
| { | |
| "epoch": 16.14, | |
| "grad_norm": 0.038977060467004776, | |
| "learning_rate": 7.446666666666668e-06, | |
| "loss": 0.0007, | |
| "step": 1650 | |
| }, | |
| { | |
| "epoch": 16.38, | |
| "grad_norm": 0.22110821306705475, | |
| "learning_rate": 7.3911111111111125e-06, | |
| "loss": 0.0007, | |
| "step": 1675 | |
| }, | |
| { | |
| "epoch": 16.63, | |
| "grad_norm": 0.5320185422897339, | |
| "learning_rate": 7.335555555555556e-06, | |
| "loss": 0.0007, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 16.87, | |
| "grad_norm": 0.7823454737663269, | |
| "learning_rate": 7.280000000000001e-06, | |
| "loss": 0.0008, | |
| "step": 1725 | |
| }, | |
| { | |
| "epoch": 17.11, | |
| "grad_norm": 0.043301377445459366, | |
| "learning_rate": 7.224444444444445e-06, | |
| "loss": 0.001, | |
| "step": 1750 | |
| }, | |
| { | |
| "epoch": 17.36, | |
| "grad_norm": 0.06231601908802986, | |
| "learning_rate": 7.1688888888888895e-06, | |
| "loss": 0.0005, | |
| "step": 1775 | |
| }, | |
| { | |
| "epoch": 17.6, | |
| "grad_norm": 0.05838339775800705, | |
| "learning_rate": 7.113333333333334e-06, | |
| "loss": 0.0005, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 17.85, | |
| "grad_norm": 0.05545497685670853, | |
| "learning_rate": 7.057777777777778e-06, | |
| "loss": 0.0012, | |
| "step": 1825 | |
| }, | |
| { | |
| "epoch": 18.09, | |
| "grad_norm": 0.4030478894710541, | |
| "learning_rate": 7.0022222222222225e-06, | |
| "loss": 0.0008, | |
| "step": 1850 | |
| }, | |
| { | |
| "epoch": 18.34, | |
| "grad_norm": 0.27439093589782715, | |
| "learning_rate": 6.946666666666667e-06, | |
| "loss": 0.0007, | |
| "step": 1875 | |
| }, | |
| { | |
| "epoch": 18.58, | |
| "grad_norm": 0.25452977418899536, | |
| "learning_rate": 6.891111111111111e-06, | |
| "loss": 0.0007, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 18.83, | |
| "grad_norm": 0.06759922206401825, | |
| "learning_rate": 6.835555555555556e-06, | |
| "loss": 0.0007, | |
| "step": 1925 | |
| }, | |
| { | |
| "epoch": 19.07, | |
| "grad_norm": 0.25859466195106506, | |
| "learning_rate": 6.780000000000001e-06, | |
| "loss": 0.0006, | |
| "step": 1950 | |
| }, | |
| { | |
| "epoch": 19.32, | |
| "grad_norm": 0.7427995800971985, | |
| "learning_rate": 6.724444444444444e-06, | |
| "loss": 0.0005, | |
| "step": 1975 | |
| }, | |
| { | |
| "epoch": 19.56, | |
| "grad_norm": 0.0788324698805809, | |
| "learning_rate": 6.668888888888889e-06, | |
| "loss": 0.0008, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 19.56, | |
| "eval_loss": 0.49481505155563354, | |
| "eval_runtime": 1457.5675, | |
| "eval_samples_per_second": 1.985, | |
| "eval_steps_per_second": 0.497, | |
| "eval_wer": 18.4735830052193, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 19.8, | |
| "grad_norm": 0.04227956011891365, | |
| "learning_rate": 6.613333333333334e-06, | |
| "loss": 0.0008, | |
| "step": 2025 | |
| }, | |
| { | |
| "epoch": 20.05, | |
| "grad_norm": 0.5580443739891052, | |
| "learning_rate": 6.557777777777778e-06, | |
| "loss": 0.001, | |
| "step": 2050 | |
| }, | |
| { | |
| "epoch": 20.29, | |
| "grad_norm": 0.7394335865974426, | |
| "learning_rate": 6.502222222222223e-06, | |
| "loss": 0.0014, | |
| "step": 2075 | |
| }, | |
| { | |
| "epoch": 20.54, | |
| "grad_norm": 0.8055688142776489, | |
| "learning_rate": 6.446666666666668e-06, | |
| "loss": 0.0011, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 20.78, | |
| "grad_norm": 0.13119255006313324, | |
| "learning_rate": 6.391111111111111e-06, | |
| "loss": 0.0016, | |
| "step": 2125 | |
| }, | |
| { | |
| "epoch": 21.03, | |
| "grad_norm": 0.21813702583312988, | |
| "learning_rate": 6.335555555555556e-06, | |
| "loss": 0.0014, | |
| "step": 2150 | |
| }, | |
| { | |
| "epoch": 21.27, | |
| "grad_norm": 0.1066213995218277, | |
| "learning_rate": 6.280000000000001e-06, | |
| "loss": 0.0009, | |
| "step": 2175 | |
| }, | |
| { | |
| "epoch": 21.52, | |
| "grad_norm": 0.8583650588989258, | |
| "learning_rate": 6.224444444444445e-06, | |
| "loss": 0.0012, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 21.76, | |
| "grad_norm": 1.2513171434402466, | |
| "learning_rate": 6.16888888888889e-06, | |
| "loss": 0.0021, | |
| "step": 2225 | |
| }, | |
| { | |
| "epoch": 22.0, | |
| "grad_norm": 0.8390223979949951, | |
| "learning_rate": 6.113333333333333e-06, | |
| "loss": 0.0018, | |
| "step": 2250 | |
| }, | |
| { | |
| "epoch": 22.25, | |
| "grad_norm": 0.8746078610420227, | |
| "learning_rate": 6.057777777777778e-06, | |
| "loss": 0.0015, | |
| "step": 2275 | |
| }, | |
| { | |
| "epoch": 22.49, | |
| "grad_norm": 0.13358770310878754, | |
| "learning_rate": 6.002222222222223e-06, | |
| "loss": 0.0016, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 22.74, | |
| "grad_norm": 0.07681471109390259, | |
| "learning_rate": 5.946666666666668e-06, | |
| "loss": 0.0011, | |
| "step": 2325 | |
| }, | |
| { | |
| "epoch": 22.98, | |
| "grad_norm": 0.5511406660079956, | |
| "learning_rate": 5.891111111111112e-06, | |
| "loss": 0.0013, | |
| "step": 2350 | |
| }, | |
| { | |
| "epoch": 23.23, | |
| "grad_norm": 0.23318354785442352, | |
| "learning_rate": 5.8355555555555565e-06, | |
| "loss": 0.0011, | |
| "step": 2375 | |
| }, | |
| { | |
| "epoch": 23.47, | |
| "grad_norm": 0.12396834790706635, | |
| "learning_rate": 5.78e-06, | |
| "loss": 0.0009, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 23.72, | |
| "grad_norm": 0.0644838809967041, | |
| "learning_rate": 5.724444444444445e-06, | |
| "loss": 0.0006, | |
| "step": 2425 | |
| }, | |
| { | |
| "epoch": 23.96, | |
| "grad_norm": 0.47172439098358154, | |
| "learning_rate": 5.6688888888888895e-06, | |
| "loss": 0.0007, | |
| "step": 2450 | |
| }, | |
| { | |
| "epoch": 24.21, | |
| "grad_norm": 0.030231019482016563, | |
| "learning_rate": 5.613333333333334e-06, | |
| "loss": 0.0005, | |
| "step": 2475 | |
| }, | |
| { | |
| "epoch": 24.45, | |
| "grad_norm": 0.01894545555114746, | |
| "learning_rate": 5.557777777777778e-06, | |
| "loss": 0.0003, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 24.69, | |
| "grad_norm": 0.18070322275161743, | |
| "learning_rate": 5.5022222222222224e-06, | |
| "loss": 0.0009, | |
| "step": 2525 | |
| }, | |
| { | |
| "epoch": 24.94, | |
| "grad_norm": 0.0435708686709404, | |
| "learning_rate": 5.4466666666666665e-06, | |
| "loss": 0.0009, | |
| "step": 2550 | |
| }, | |
| { | |
| "epoch": 25.18, | |
| "grad_norm": 0.07220447063446045, | |
| "learning_rate": 5.391111111111111e-06, | |
| "loss": 0.0005, | |
| "step": 2575 | |
| }, | |
| { | |
| "epoch": 25.43, | |
| "grad_norm": 0.01733986660838127, | |
| "learning_rate": 5.335555555555556e-06, | |
| "loss": 0.0004, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 25.67, | |
| "grad_norm": 0.03520004078745842, | |
| "learning_rate": 5.28e-06, | |
| "loss": 0.0007, | |
| "step": 2625 | |
| }, | |
| { | |
| "epoch": 25.92, | |
| "grad_norm": 0.03853292763233185, | |
| "learning_rate": 5.224444444444445e-06, | |
| "loss": 0.0005, | |
| "step": 2650 | |
| }, | |
| { | |
| "epoch": 26.16, | |
| "grad_norm": 0.13450591266155243, | |
| "learning_rate": 5.168888888888889e-06, | |
| "loss": 0.0005, | |
| "step": 2675 | |
| }, | |
| { | |
| "epoch": 26.41, | |
| "grad_norm": 0.029255390167236328, | |
| "learning_rate": 5.113333333333333e-06, | |
| "loss": 0.0004, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 26.65, | |
| "grad_norm": 0.025706447660923004, | |
| "learning_rate": 5.057777777777778e-06, | |
| "loss": 0.0003, | |
| "step": 2725 | |
| }, | |
| { | |
| "epoch": 26.89, | |
| "grad_norm": 0.902415931224823, | |
| "learning_rate": 5.002222222222223e-06, | |
| "loss": 0.0002, | |
| "step": 2750 | |
| }, | |
| { | |
| "epoch": 27.14, | |
| "grad_norm": 0.013656423427164555, | |
| "learning_rate": 4.946666666666667e-06, | |
| "loss": 0.0002, | |
| "step": 2775 | |
| }, | |
| { | |
| "epoch": 27.38, | |
| "grad_norm": 0.018052740022540092, | |
| "learning_rate": 4.891111111111111e-06, | |
| "loss": 0.0003, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 27.63, | |
| "grad_norm": 0.07001502811908722, | |
| "learning_rate": 4.835555555555556e-06, | |
| "loss": 0.0002, | |
| "step": 2825 | |
| }, | |
| { | |
| "epoch": 27.87, | |
| "grad_norm": 0.01241993810981512, | |
| "learning_rate": 4.78e-06, | |
| "loss": 0.0003, | |
| "step": 2850 | |
| }, | |
| { | |
| "epoch": 28.12, | |
| "grad_norm": 0.013495378196239471, | |
| "learning_rate": 4.724444444444445e-06, | |
| "loss": 0.0001, | |
| "step": 2875 | |
| }, | |
| { | |
| "epoch": 28.36, | |
| "grad_norm": 0.011048069223761559, | |
| "learning_rate": 4.66888888888889e-06, | |
| "loss": 0.0001, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 28.61, | |
| "grad_norm": 0.015970442444086075, | |
| "learning_rate": 4.613333333333334e-06, | |
| "loss": 0.0002, | |
| "step": 2925 | |
| }, | |
| { | |
| "epoch": 28.85, | |
| "grad_norm": 0.009559527039527893, | |
| "learning_rate": 4.557777777777778e-06, | |
| "loss": 0.0002, | |
| "step": 2950 | |
| }, | |
| { | |
| "epoch": 29.1, | |
| "grad_norm": 0.010293275117874146, | |
| "learning_rate": 4.502222222222223e-06, | |
| "loss": 0.0001, | |
| "step": 2975 | |
| }, | |
| { | |
| "epoch": 29.34, | |
| "grad_norm": 0.01116804126650095, | |
| "learning_rate": 4.446666666666667e-06, | |
| "loss": 0.0001, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 29.34, | |
| "eval_loss": 0.5352661609649658, | |
| "eval_runtime": 1458.2257, | |
| "eval_samples_per_second": 1.985, | |
| "eval_steps_per_second": 0.496, | |
| "eval_wer": 18.07297866495742, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 29.58, | |
| "grad_norm": 0.0082534896209836, | |
| "learning_rate": 4.391111111111112e-06, | |
| "loss": 0.0001, | |
| "step": 3025 | |
| }, | |
| { | |
| "epoch": 29.83, | |
| "grad_norm": 0.007231541443616152, | |
| "learning_rate": 4.3355555555555565e-06, | |
| "loss": 0.0001, | |
| "step": 3050 | |
| }, | |
| { | |
| "epoch": 30.07, | |
| "grad_norm": 0.008312270976603031, | |
| "learning_rate": 4.2800000000000005e-06, | |
| "loss": 0.0001, | |
| "step": 3075 | |
| }, | |
| { | |
| "epoch": 30.32, | |
| "grad_norm": 0.007097834721207619, | |
| "learning_rate": 4.2244444444444446e-06, | |
| "loss": 0.0001, | |
| "step": 3100 | |
| }, | |
| { | |
| "epoch": 30.56, | |
| "grad_norm": 0.007524041458964348, | |
| "learning_rate": 4.168888888888889e-06, | |
| "loss": 0.0001, | |
| "step": 3125 | |
| }, | |
| { | |
| "epoch": 30.81, | |
| "grad_norm": 0.00781218009069562, | |
| "learning_rate": 4.1133333333333335e-06, | |
| "loss": 0.0001, | |
| "step": 3150 | |
| }, | |
| { | |
| "epoch": 31.05, | |
| "grad_norm": 0.006727377884089947, | |
| "learning_rate": 4.057777777777778e-06, | |
| "loss": 0.0001, | |
| "step": 3175 | |
| }, | |
| { | |
| "epoch": 31.3, | |
| "grad_norm": 0.00624061468988657, | |
| "learning_rate": 4.002222222222222e-06, | |
| "loss": 0.0001, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 31.54, | |
| "grad_norm": 0.006174057722091675, | |
| "learning_rate": 3.946666666666667e-06, | |
| "loss": 0.0001, | |
| "step": 3225 | |
| }, | |
| { | |
| "epoch": 31.78, | |
| "grad_norm": 0.006670523434877396, | |
| "learning_rate": 3.891111111111111e-06, | |
| "loss": 0.0001, | |
| "step": 3250 | |
| }, | |
| { | |
| "epoch": 32.03, | |
| "grad_norm": 0.006069181486964226, | |
| "learning_rate": 3.835555555555555e-06, | |
| "loss": 0.0001, | |
| "step": 3275 | |
| }, | |
| { | |
| "epoch": 32.27, | |
| "grad_norm": 0.005773240700364113, | |
| "learning_rate": 3.7800000000000002e-06, | |
| "loss": 0.0001, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 32.52, | |
| "grad_norm": 0.005736664403229952, | |
| "learning_rate": 3.724444444444445e-06, | |
| "loss": 0.0001, | |
| "step": 3325 | |
| }, | |
| { | |
| "epoch": 32.76, | |
| "grad_norm": 0.0057275379076600075, | |
| "learning_rate": 3.668888888888889e-06, | |
| "loss": 0.0001, | |
| "step": 3350 | |
| }, | |
| { | |
| "epoch": 33.01, | |
| "grad_norm": 0.0070615834556519985, | |
| "learning_rate": 3.6133333333333336e-06, | |
| "loss": 0.0001, | |
| "step": 3375 | |
| }, | |
| { | |
| "epoch": 33.25, | |
| "grad_norm": 0.005553886294364929, | |
| "learning_rate": 3.5577777777777785e-06, | |
| "loss": 0.0001, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 33.5, | |
| "grad_norm": 0.00474073551595211, | |
| "learning_rate": 3.5022222222222225e-06, | |
| "loss": 0.0001, | |
| "step": 3425 | |
| }, | |
| { | |
| "epoch": 33.74, | |
| "grad_norm": 0.006729442626237869, | |
| "learning_rate": 3.446666666666667e-06, | |
| "loss": 0.0001, | |
| "step": 3450 | |
| }, | |
| { | |
| "epoch": 33.99, | |
| "grad_norm": 0.006181245669722557, | |
| "learning_rate": 3.391111111111111e-06, | |
| "loss": 0.0001, | |
| "step": 3475 | |
| }, | |
| { | |
| "epoch": 34.23, | |
| "grad_norm": 0.004282401409000158, | |
| "learning_rate": 3.335555555555556e-06, | |
| "loss": 0.0001, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 34.47, | |
| "grad_norm": 0.0049114222638309, | |
| "learning_rate": 3.2800000000000004e-06, | |
| "loss": 0.0001, | |
| "step": 3525 | |
| }, | |
| { | |
| "epoch": 34.72, | |
| "grad_norm": 0.004997397307306528, | |
| "learning_rate": 3.2244444444444444e-06, | |
| "loss": 0.0001, | |
| "step": 3550 | |
| }, | |
| { | |
| "epoch": 34.96, | |
| "grad_norm": 0.004732249770313501, | |
| "learning_rate": 3.1688888888888893e-06, | |
| "loss": 0.0001, | |
| "step": 3575 | |
| }, | |
| { | |
| "epoch": 35.21, | |
| "grad_norm": 0.005105483811348677, | |
| "learning_rate": 3.1133333333333337e-06, | |
| "loss": 0.0001, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 35.45, | |
| "grad_norm": 0.005223471205681562, | |
| "learning_rate": 3.0577777777777778e-06, | |
| "loss": 0.0001, | |
| "step": 3625 | |
| }, | |
| { | |
| "epoch": 35.7, | |
| "grad_norm": 0.005028573330491781, | |
| "learning_rate": 3.0022222222222227e-06, | |
| "loss": 0.0001, | |
| "step": 3650 | |
| }, | |
| { | |
| "epoch": 35.94, | |
| "grad_norm": 0.0035760572645813227, | |
| "learning_rate": 2.946666666666667e-06, | |
| "loss": 0.0001, | |
| "step": 3675 | |
| }, | |
| { | |
| "epoch": 36.19, | |
| "grad_norm": 0.0042958687990903854, | |
| "learning_rate": 2.891111111111111e-06, | |
| "loss": 0.0001, | |
| "step": 3700 | |
| }, | |
| { | |
| "epoch": 36.43, | |
| "grad_norm": 0.004194607958197594, | |
| "learning_rate": 2.835555555555556e-06, | |
| "loss": 0.0001, | |
| "step": 3725 | |
| }, | |
| { | |
| "epoch": 36.67, | |
| "grad_norm": 0.004419374745339155, | |
| "learning_rate": 2.7800000000000005e-06, | |
| "loss": 0.0001, | |
| "step": 3750 | |
| }, | |
| { | |
| "epoch": 36.92, | |
| "grad_norm": 0.004165703430771828, | |
| "learning_rate": 2.7244444444444445e-06, | |
| "loss": 0.0001, | |
| "step": 3775 | |
| }, | |
| { | |
| "epoch": 37.16, | |
| "grad_norm": 0.0038676238618791103, | |
| "learning_rate": 2.6688888888888894e-06, | |
| "loss": 0.0001, | |
| "step": 3800 | |
| }, | |
| { | |
| "epoch": 37.41, | |
| "grad_norm": 0.004417106043547392, | |
| "learning_rate": 2.6133333333333334e-06, | |
| "loss": 0.0001, | |
| "step": 3825 | |
| }, | |
| { | |
| "epoch": 37.65, | |
| "grad_norm": 0.004240726120769978, | |
| "learning_rate": 2.557777777777778e-06, | |
| "loss": 0.0001, | |
| "step": 3850 | |
| }, | |
| { | |
| "epoch": 37.9, | |
| "grad_norm": 0.003732978831976652, | |
| "learning_rate": 2.5022222222222224e-06, | |
| "loss": 0.0001, | |
| "step": 3875 | |
| }, | |
| { | |
| "epoch": 38.14, | |
| "grad_norm": 0.0037496890872716904, | |
| "learning_rate": 2.446666666666667e-06, | |
| "loss": 0.0001, | |
| "step": 3900 | |
| }, | |
| { | |
| "epoch": 38.39, | |
| "grad_norm": 0.003918818198144436, | |
| "learning_rate": 2.3911111111111113e-06, | |
| "loss": 0.0001, | |
| "step": 3925 | |
| }, | |
| { | |
| "epoch": 38.63, | |
| "grad_norm": 0.003961279056966305, | |
| "learning_rate": 2.3355555555555557e-06, | |
| "loss": 0.0001, | |
| "step": 3950 | |
| }, | |
| { | |
| "epoch": 38.88, | |
| "grad_norm": 0.00358415674418211, | |
| "learning_rate": 2.28e-06, | |
| "loss": 0.0001, | |
| "step": 3975 | |
| }, | |
| { | |
| "epoch": 39.12, | |
| "grad_norm": 0.004146341234445572, | |
| "learning_rate": 2.2244444444444447e-06, | |
| "loss": 0.0001, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 39.12, | |
| "eval_loss": 0.5624426603317261, | |
| "eval_runtime": 1469.1635, | |
| "eval_samples_per_second": 1.97, | |
| "eval_steps_per_second": 0.493, | |
| "eval_wer": 18.056954491346946, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 39.36, | |
| "grad_norm": 0.004992615897208452, | |
| "learning_rate": 2.168888888888889e-06, | |
| "loss": 0.0001, | |
| "step": 4025 | |
| }, | |
| { | |
| "epoch": 39.61, | |
| "grad_norm": 0.0034102711360901594, | |
| "learning_rate": 2.1133333333333336e-06, | |
| "loss": 0.0001, | |
| "step": 4050 | |
| }, | |
| { | |
| "epoch": 39.85, | |
| "grad_norm": 0.0036119353026151657, | |
| "learning_rate": 2.057777777777778e-06, | |
| "loss": 0.0, | |
| "step": 4075 | |
| }, | |
| { | |
| "epoch": 40.1, | |
| "grad_norm": 0.0040961061604321, | |
| "learning_rate": 2.0022222222222225e-06, | |
| "loss": 0.0001, | |
| "step": 4100 | |
| }, | |
| { | |
| "epoch": 40.34, | |
| "grad_norm": 0.003933845553547144, | |
| "learning_rate": 1.9466666666666665e-06, | |
| "loss": 0.0001, | |
| "step": 4125 | |
| }, | |
| { | |
| "epoch": 40.59, | |
| "grad_norm": 0.00343153509311378, | |
| "learning_rate": 1.8911111111111114e-06, | |
| "loss": 0.0, | |
| "step": 4150 | |
| }, | |
| { | |
| "epoch": 40.83, | |
| "grad_norm": 0.0032181148417294025, | |
| "learning_rate": 1.8355555555555557e-06, | |
| "loss": 0.0, | |
| "step": 4175 | |
| }, | |
| { | |
| "epoch": 41.08, | |
| "grad_norm": 0.002862308407202363, | |
| "learning_rate": 1.7800000000000001e-06, | |
| "loss": 0.0, | |
| "step": 4200 | |
| }, | |
| { | |
| "epoch": 41.32, | |
| "grad_norm": 0.0038496414199471474, | |
| "learning_rate": 1.7244444444444448e-06, | |
| "loss": 0.0, | |
| "step": 4225 | |
| }, | |
| { | |
| "epoch": 41.56, | |
| "grad_norm": 0.0036370421294122934, | |
| "learning_rate": 1.668888888888889e-06, | |
| "loss": 0.0, | |
| "step": 4250 | |
| }, | |
| { | |
| "epoch": 41.81, | |
| "grad_norm": 0.003030687803402543, | |
| "learning_rate": 1.6133333333333335e-06, | |
| "loss": 0.0, | |
| "step": 4275 | |
| }, | |
| { | |
| "epoch": 42.05, | |
| "grad_norm": 0.003104611998423934, | |
| "learning_rate": 1.5577777777777777e-06, | |
| "loss": 0.0, | |
| "step": 4300 | |
| }, | |
| { | |
| "epoch": 42.3, | |
| "grad_norm": 0.0035941856913268566, | |
| "learning_rate": 1.5022222222222224e-06, | |
| "loss": 0.0, | |
| "step": 4325 | |
| }, | |
| { | |
| "epoch": 42.54, | |
| "grad_norm": 0.0029717443976551294, | |
| "learning_rate": 1.4466666666666669e-06, | |
| "loss": 0.0, | |
| "step": 4350 | |
| }, | |
| { | |
| "epoch": 42.79, | |
| "grad_norm": 0.0039994968101382256, | |
| "learning_rate": 1.3911111111111111e-06, | |
| "loss": 0.0, | |
| "step": 4375 | |
| }, | |
| { | |
| "epoch": 43.03, | |
| "grad_norm": 0.003351717721670866, | |
| "learning_rate": 1.3355555555555558e-06, | |
| "loss": 0.0, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 43.28, | |
| "grad_norm": 0.00382447661831975, | |
| "learning_rate": 1.28e-06, | |
| "loss": 0.0, | |
| "step": 4425 | |
| }, | |
| { | |
| "epoch": 43.52, | |
| "grad_norm": 0.0030940717551857233, | |
| "learning_rate": 1.2244444444444445e-06, | |
| "loss": 0.0, | |
| "step": 4450 | |
| }, | |
| { | |
| "epoch": 43.77, | |
| "grad_norm": 0.003348903963342309, | |
| "learning_rate": 1.168888888888889e-06, | |
| "loss": 0.0, | |
| "step": 4475 | |
| }, | |
| { | |
| "epoch": 44.01, | |
| "grad_norm": 0.002947325585409999, | |
| "learning_rate": 1.1133333333333334e-06, | |
| "loss": 0.0, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 44.25, | |
| "grad_norm": 0.0035392455756664276, | |
| "learning_rate": 1.0577777777777779e-06, | |
| "loss": 0.0, | |
| "step": 4525 | |
| }, | |
| { | |
| "epoch": 44.5, | |
| "grad_norm": 0.0031909053213894367, | |
| "learning_rate": 1.0022222222222223e-06, | |
| "loss": 0.0, | |
| "step": 4550 | |
| }, | |
| { | |
| "epoch": 44.74, | |
| "grad_norm": 0.0031825106125324965, | |
| "learning_rate": 9.466666666666667e-07, | |
| "loss": 0.0, | |
| "step": 4575 | |
| }, | |
| { | |
| "epoch": 44.99, | |
| "grad_norm": 0.003659332636743784, | |
| "learning_rate": 8.911111111111112e-07, | |
| "loss": 0.0, | |
| "step": 4600 | |
| }, | |
| { | |
| "epoch": 45.23, | |
| "grad_norm": 0.003141431836411357, | |
| "learning_rate": 8.355555555555556e-07, | |
| "loss": 0.0, | |
| "step": 4625 | |
| }, | |
| { | |
| "epoch": 45.48, | |
| "grad_norm": 0.0026865031104534864, | |
| "learning_rate": 7.8e-07, | |
| "loss": 0.0, | |
| "step": 4650 | |
| }, | |
| { | |
| "epoch": 45.72, | |
| "grad_norm": 0.0037062685005366802, | |
| "learning_rate": 7.244444444444446e-07, | |
| "loss": 0.0, | |
| "step": 4675 | |
| }, | |
| { | |
| "epoch": 45.97, | |
| "grad_norm": 0.0038293024990707636, | |
| "learning_rate": 6.68888888888889e-07, | |
| "loss": 0.0, | |
| "step": 4700 | |
| }, | |
| { | |
| "epoch": 46.21, | |
| "grad_norm": 0.0033506678882986307, | |
| "learning_rate": 6.133333333333333e-07, | |
| "loss": 0.0, | |
| "step": 4725 | |
| }, | |
| { | |
| "epoch": 46.45, | |
| "grad_norm": 0.0031830337829887867, | |
| "learning_rate": 5.577777777777779e-07, | |
| "loss": 0.0, | |
| "step": 4750 | |
| }, | |
| { | |
| "epoch": 46.7, | |
| "grad_norm": 0.002798918168991804, | |
| "learning_rate": 5.022222222222222e-07, | |
| "loss": 0.0, | |
| "step": 4775 | |
| }, | |
| { | |
| "epoch": 46.94, | |
| "grad_norm": 0.0032959673553705215, | |
| "learning_rate": 4.466666666666667e-07, | |
| "loss": 0.0, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 47.19, | |
| "grad_norm": 0.0029275265987962484, | |
| "learning_rate": 3.9111111111111115e-07, | |
| "loss": 0.0, | |
| "step": 4825 | |
| }, | |
| { | |
| "epoch": 47.43, | |
| "grad_norm": 0.0030392766930162907, | |
| "learning_rate": 3.3555555555555556e-07, | |
| "loss": 0.0, | |
| "step": 4850 | |
| }, | |
| { | |
| "epoch": 47.68, | |
| "grad_norm": 0.002865479327738285, | |
| "learning_rate": 2.8e-07, | |
| "loss": 0.0, | |
| "step": 4875 | |
| }, | |
| { | |
| "epoch": 47.92, | |
| "grad_norm": 0.0030035064555704594, | |
| "learning_rate": 2.2444444444444445e-07, | |
| "loss": 0.0, | |
| "step": 4900 | |
| }, | |
| { | |
| "epoch": 48.17, | |
| "grad_norm": 0.00276967347599566, | |
| "learning_rate": 1.6888888888888888e-07, | |
| "loss": 0.0, | |
| "step": 4925 | |
| }, | |
| { | |
| "epoch": 48.41, | |
| "grad_norm": 0.003191766096279025, | |
| "learning_rate": 1.1333333333333336e-07, | |
| "loss": 0.0, | |
| "step": 4950 | |
| }, | |
| { | |
| "epoch": 48.66, | |
| "grad_norm": 0.00321377394720912, | |
| "learning_rate": 5.777777777777778e-08, | |
| "loss": 0.0, | |
| "step": 4975 | |
| }, | |
| { | |
| "epoch": 48.9, | |
| "grad_norm": 0.0033689856063574553, | |
| "learning_rate": 2.2222222222222225e-09, | |
| "loss": 0.0, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 48.9, | |
| "eval_loss": 0.5714073181152344, | |
| "eval_runtime": 1474.141, | |
| "eval_samples_per_second": 1.963, | |
| "eval_steps_per_second": 0.491, | |
| "eval_wer": 18.063821994322865, | |
| "step": 5000 | |
| } | |
| ], | |
| "logging_steps": 25, | |
| "max_steps": 5000, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 50, | |
| "save_steps": 1000, | |
| "total_flos": 9.229191970553856e+19, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |