| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.2825445684237526, | |
| "eval_steps": 500, | |
| "global_step": 20000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.03206361421059382, | |
| "grad_norm": 9.577472686767578, | |
| "learning_rate": 0.0002465, | |
| "loss": 4.7126, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.03206361421059382, | |
| "eval_loss": 1.7046922445297241, | |
| "eval_runtime": 189.0412, | |
| "eval_samples_per_second": 37.145, | |
| "eval_steps_per_second": 0.582, | |
| "eval_wer": 0.9345593218086873, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.06412722842118763, | |
| "grad_norm": 6.515851020812988, | |
| "learning_rate": 0.0002939226804123711, | |
| "loss": 1.0533, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.06412722842118763, | |
| "eval_loss": 1.1487088203430176, | |
| "eval_runtime": 189.0728, | |
| "eval_samples_per_second": 37.139, | |
| "eval_steps_per_second": 0.582, | |
| "eval_wer": 0.7906695544718904, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.09619084263178146, | |
| "grad_norm": 3.676572799682617, | |
| "learning_rate": 0.0002861907216494845, | |
| "loss": 0.8268, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.09619084263178146, | |
| "eval_loss": 1.060188889503479, | |
| "eval_runtime": 190.9733, | |
| "eval_samples_per_second": 36.77, | |
| "eval_steps_per_second": 0.576, | |
| "eval_wer": 0.7815012158014324, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.12825445684237527, | |
| "grad_norm": 9.430739402770996, | |
| "learning_rate": 0.00027845876288659795, | |
| "loss": 0.7188, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.12825445684237527, | |
| "eval_loss": 0.9336337447166443, | |
| "eval_runtime": 191.0896, | |
| "eval_samples_per_second": 36.747, | |
| "eval_steps_per_second": 0.576, | |
| "eval_wer": 0.671746900702906, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.16031807105296908, | |
| "grad_norm": 2.9828200340270996, | |
| "learning_rate": 0.00027072680412371135, | |
| "loss": 0.6725, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.16031807105296908, | |
| "eval_loss": 0.9303568005561829, | |
| "eval_runtime": 191.2157, | |
| "eval_samples_per_second": 36.723, | |
| "eval_steps_per_second": 0.575, | |
| "eval_wer": 0.6560677128316879, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.19238168526356292, | |
| "grad_norm": 4.710850238800049, | |
| "learning_rate": 0.0002629948453608247, | |
| "loss": 0.6295, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.19238168526356292, | |
| "eval_loss": 0.8600214719772339, | |
| "eval_runtime": 191.6797, | |
| "eval_samples_per_second": 36.634, | |
| "eval_steps_per_second": 0.574, | |
| "eval_wer": 0.6257324705350855, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.22444529947415673, | |
| "grad_norm": 4.912868976593018, | |
| "learning_rate": 0.0002552628865979381, | |
| "loss": 0.6003, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 0.22444529947415673, | |
| "eval_loss": 0.8395254611968994, | |
| "eval_runtime": 191.3108, | |
| "eval_samples_per_second": 36.705, | |
| "eval_steps_per_second": 0.575, | |
| "eval_wer": 0.6113288776093224, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 0.25650891368475054, | |
| "grad_norm": 4.513955116271973, | |
| "learning_rate": 0.00024754639175257734, | |
| "loss": 0.5847, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.25650891368475054, | |
| "eval_loss": 0.7883865833282471, | |
| "eval_runtime": 192.8783, | |
| "eval_samples_per_second": 36.406, | |
| "eval_steps_per_second": 0.57, | |
| "eval_wer": 0.5861491648839341, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.2885725278953444, | |
| "grad_norm": 16.630624771118164, | |
| "learning_rate": 0.00023984536082474227, | |
| "loss": 0.5521, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 0.2885725278953444, | |
| "eval_loss": 0.7741186618804932, | |
| "eval_runtime": 189.6516, | |
| "eval_samples_per_second": 37.026, | |
| "eval_steps_per_second": 0.58, | |
| "eval_wer": 0.5686628841733214, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 0.32063614210593816, | |
| "grad_norm": 7.58245325088501, | |
| "learning_rate": 0.00023211340206185567, | |
| "loss": 0.5477, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.32063614210593816, | |
| "eval_loss": 0.7594121098518372, | |
| "eval_runtime": 190.5466, | |
| "eval_samples_per_second": 36.852, | |
| "eval_steps_per_second": 0.577, | |
| "eval_wer": 0.5535550565380885, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.352699756316532, | |
| "grad_norm": 5.051167011260986, | |
| "learning_rate": 0.00022438144329896904, | |
| "loss": 0.5346, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 0.352699756316532, | |
| "eval_loss": 0.7481973767280579, | |
| "eval_runtime": 190.5699, | |
| "eval_samples_per_second": 36.847, | |
| "eval_steps_per_second": 0.577, | |
| "eval_wer": 0.5394039251119468, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 0.38476337052712584, | |
| "grad_norm": 4.212076187133789, | |
| "learning_rate": 0.00021666494845360825, | |
| "loss": 0.5154, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.38476337052712584, | |
| "eval_loss": 0.7294158935546875, | |
| "eval_runtime": 189.7232, | |
| "eval_samples_per_second": 37.012, | |
| "eval_steps_per_second": 0.58, | |
| "eval_wer": 0.53515194196043, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.4168269847377196, | |
| "grad_norm": 5.682095527648926, | |
| "learning_rate": 0.00020893298969072165, | |
| "loss": 0.492, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 0.4168269847377196, | |
| "eval_loss": 0.7247592806816101, | |
| "eval_runtime": 190.6553, | |
| "eval_samples_per_second": 36.831, | |
| "eval_steps_per_second": 0.577, | |
| "eval_wer": 0.5492632110445262, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 0.44889059894831346, | |
| "grad_norm": 8.364203453063965, | |
| "learning_rate": 0.0002012164948453608, | |
| "loss": 0.4759, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 0.44889059894831346, | |
| "eval_loss": 0.7076719403266907, | |
| "eval_runtime": 189.5572, | |
| "eval_samples_per_second": 37.044, | |
| "eval_steps_per_second": 0.58, | |
| "eval_wer": 0.5134402529929976, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 0.4809542131589073, | |
| "grad_norm": 4.447290897369385, | |
| "learning_rate": 0.0001934845360824742, | |
| "loss": 0.4655, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 0.4809542131589073, | |
| "eval_loss": 0.673875629901886, | |
| "eval_runtime": 190.3324, | |
| "eval_samples_per_second": 36.893, | |
| "eval_steps_per_second": 0.578, | |
| "eval_wer": 0.5063979058982979, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 0.5130178273695011, | |
| "grad_norm": 12.618865013122559, | |
| "learning_rate": 0.0001857680412371134, | |
| "loss": 0.4594, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 0.5130178273695011, | |
| "eval_loss": 0.6574720144271851, | |
| "eval_runtime": 190.8303, | |
| "eval_samples_per_second": 36.797, | |
| "eval_steps_per_second": 0.576, | |
| "eval_wer": 0.5067300920820101, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 0.5450814415800949, | |
| "grad_norm": 2.756011962890625, | |
| "learning_rate": 0.0001780360824742268, | |
| "loss": 0.4538, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 0.5450814415800949, | |
| "eval_loss": 0.6492609977722168, | |
| "eval_runtime": 189.5472, | |
| "eval_samples_per_second": 37.046, | |
| "eval_steps_per_second": 0.58, | |
| "eval_wer": 0.500325542460038, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 0.5771450557906888, | |
| "grad_norm": 7.6861491203308105, | |
| "learning_rate": 0.0001703041237113402, | |
| "loss": 0.4739, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 0.5771450557906888, | |
| "eval_loss": 0.7676782608032227, | |
| "eval_runtime": 189.9033, | |
| "eval_samples_per_second": 36.977, | |
| "eval_steps_per_second": 0.579, | |
| "eval_wer": 0.5238576117142136, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 0.6092086700012825, | |
| "grad_norm": NaN, | |
| "learning_rate": 0.00016787628865979378, | |
| "loss": 0.695, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 0.6092086700012825, | |
| "eval_loss": NaN, | |
| "eval_runtime": 186.9158, | |
| "eval_samples_per_second": 37.568, | |
| "eval_steps_per_second": 0.589, | |
| "eval_wer": 1.0, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 0.6412722842118763, | |
| "grad_norm": NaN, | |
| "learning_rate": 0.00016787628865979378, | |
| "loss": 0.0, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 0.6412722842118763, | |
| "eval_loss": NaN, | |
| "eval_runtime": 186.7613, | |
| "eval_samples_per_second": 37.599, | |
| "eval_steps_per_second": 0.589, | |
| "eval_wer": 1.0, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 0.6733358984224702, | |
| "grad_norm": NaN, | |
| "learning_rate": 0.00016787628865979378, | |
| "loss": 0.0, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 0.6733358984224702, | |
| "eval_loss": NaN, | |
| "eval_runtime": 185.9746, | |
| "eval_samples_per_second": 37.758, | |
| "eval_steps_per_second": 0.591, | |
| "eval_wer": 1.0, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 0.705399512633064, | |
| "grad_norm": NaN, | |
| "learning_rate": 0.00016787628865979378, | |
| "loss": 0.0, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 0.705399512633064, | |
| "eval_loss": NaN, | |
| "eval_runtime": 186.0588, | |
| "eval_samples_per_second": 37.741, | |
| "eval_steps_per_second": 0.591, | |
| "eval_wer": 1.0, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 0.7374631268436578, | |
| "grad_norm": NaN, | |
| "learning_rate": 0.00016787628865979378, | |
| "loss": 0.0, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 0.7374631268436578, | |
| "eval_loss": NaN, | |
| "eval_runtime": 187.0825, | |
| "eval_samples_per_second": 37.534, | |
| "eval_steps_per_second": 0.588, | |
| "eval_wer": 1.0, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 0.7695267410542517, | |
| "grad_norm": NaN, | |
| "learning_rate": 0.00016787628865979378, | |
| "loss": 0.0, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 0.7695267410542517, | |
| "eval_loss": NaN, | |
| "eval_runtime": 186.5183, | |
| "eval_samples_per_second": 37.648, | |
| "eval_steps_per_second": 0.59, | |
| "eval_wer": 1.0, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 0.8015903552648455, | |
| "grad_norm": NaN, | |
| "learning_rate": 0.00016787628865979378, | |
| "loss": 0.0, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 0.8015903552648455, | |
| "eval_loss": NaN, | |
| "eval_runtime": 186.3281, | |
| "eval_samples_per_second": 37.686, | |
| "eval_steps_per_second": 0.59, | |
| "eval_wer": 1.0, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 0.8336539694754392, | |
| "grad_norm": NaN, | |
| "learning_rate": 0.00016787628865979378, | |
| "loss": 0.0, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 0.8336539694754392, | |
| "eval_loss": NaN, | |
| "eval_runtime": 185.5922, | |
| "eval_samples_per_second": 37.836, | |
| "eval_steps_per_second": 0.593, | |
| "eval_wer": 1.0, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 0.8657175836860331, | |
| "grad_norm": NaN, | |
| "learning_rate": 0.00016787628865979378, | |
| "loss": 0.0, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 0.8657175836860331, | |
| "eval_loss": NaN, | |
| "eval_runtime": 185.7237, | |
| "eval_samples_per_second": 37.809, | |
| "eval_steps_per_second": 0.592, | |
| "eval_wer": 1.0, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 0.8977811978966269, | |
| "grad_norm": NaN, | |
| "learning_rate": 0.00016787628865979378, | |
| "loss": 0.0, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 0.8977811978966269, | |
| "eval_loss": NaN, | |
| "eval_runtime": 186.6259, | |
| "eval_samples_per_second": 37.626, | |
| "eval_steps_per_second": 0.589, | |
| "eval_wer": 1.0, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 0.9298448121072207, | |
| "grad_norm": NaN, | |
| "learning_rate": 0.00016787628865979378, | |
| "loss": 0.0, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 0.9298448121072207, | |
| "eval_loss": NaN, | |
| "eval_runtime": 186.1517, | |
| "eval_samples_per_second": 37.722, | |
| "eval_steps_per_second": 0.591, | |
| "eval_wer": 1.0, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 0.9619084263178146, | |
| "grad_norm": NaN, | |
| "learning_rate": 0.00016787628865979378, | |
| "loss": 0.0, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 0.9619084263178146, | |
| "eval_loss": NaN, | |
| "eval_runtime": 186.7927, | |
| "eval_samples_per_second": 37.592, | |
| "eval_steps_per_second": 0.589, | |
| "eval_wer": 1.0, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 0.9939720405284084, | |
| "grad_norm": NaN, | |
| "learning_rate": 0.00016787628865979378, | |
| "loss": 0.0, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 0.9939720405284084, | |
| "eval_loss": NaN, | |
| "eval_runtime": 186.1708, | |
| "eval_samples_per_second": 37.718, | |
| "eval_steps_per_second": 0.591, | |
| "eval_wer": 1.0, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 1.0260356547390022, | |
| "grad_norm": NaN, | |
| "learning_rate": 0.00016787628865979378, | |
| "loss": 0.0, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 1.0260356547390022, | |
| "eval_loss": NaN, | |
| "eval_runtime": 186.1341, | |
| "eval_samples_per_second": 37.725, | |
| "eval_steps_per_second": 0.591, | |
| "eval_wer": 1.0, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 1.058099268949596, | |
| "grad_norm": NaN, | |
| "learning_rate": 0.00016787628865979378, | |
| "loss": 0.0, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 1.058099268949596, | |
| "eval_loss": NaN, | |
| "eval_runtime": 186.4575, | |
| "eval_samples_per_second": 37.66, | |
| "eval_steps_per_second": 0.59, | |
| "eval_wer": 1.0, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 1.0901628831601897, | |
| "grad_norm": NaN, | |
| "learning_rate": 0.00016787628865979378, | |
| "loss": 0.0, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 1.0901628831601897, | |
| "eval_loss": NaN, | |
| "eval_runtime": 185.4444, | |
| "eval_samples_per_second": 37.866, | |
| "eval_steps_per_second": 0.593, | |
| "eval_wer": 1.0, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 1.1222264973707836, | |
| "grad_norm": NaN, | |
| "learning_rate": 0.00016787628865979378, | |
| "loss": 0.0, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 1.1222264973707836, | |
| "eval_loss": NaN, | |
| "eval_runtime": 186.15, | |
| "eval_samples_per_second": 37.722, | |
| "eval_steps_per_second": 0.591, | |
| "eval_wer": 1.0, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 1.1542901115813775, | |
| "grad_norm": NaN, | |
| "learning_rate": 0.00016787628865979378, | |
| "loss": 0.0, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 1.1542901115813775, | |
| "eval_loss": NaN, | |
| "eval_runtime": 186.0027, | |
| "eval_samples_per_second": 37.752, | |
| "eval_steps_per_second": 0.591, | |
| "eval_wer": 1.0, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 1.1863537257919712, | |
| "grad_norm": NaN, | |
| "learning_rate": 0.00016787628865979378, | |
| "loss": 0.0, | |
| "step": 18500 | |
| }, | |
| { | |
| "epoch": 1.1863537257919712, | |
| "eval_loss": NaN, | |
| "eval_runtime": 185.6149, | |
| "eval_samples_per_second": 37.831, | |
| "eval_steps_per_second": 0.593, | |
| "eval_wer": 1.0, | |
| "step": 18500 | |
| }, | |
| { | |
| "epoch": 1.218417340002565, | |
| "grad_norm": NaN, | |
| "learning_rate": 0.00016787628865979378, | |
| "loss": 0.0, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 1.218417340002565, | |
| "eval_loss": NaN, | |
| "eval_runtime": 186.7557, | |
| "eval_samples_per_second": 37.6, | |
| "eval_steps_per_second": 0.589, | |
| "eval_wer": 1.0, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 1.250480954213159, | |
| "grad_norm": NaN, | |
| "learning_rate": 0.00016787628865979378, | |
| "loss": 0.0, | |
| "step": 19500 | |
| }, | |
| { | |
| "epoch": 1.250480954213159, | |
| "eval_loss": NaN, | |
| "eval_runtime": 186.7166, | |
| "eval_samples_per_second": 37.608, | |
| "eval_steps_per_second": 0.589, | |
| "eval_wer": 1.0, | |
| "step": 19500 | |
| }, | |
| { | |
| "epoch": 1.2825445684237526, | |
| "grad_norm": NaN, | |
| "learning_rate": 0.00016787628865979378, | |
| "loss": 0.0, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 1.2825445684237526, | |
| "eval_loss": NaN, | |
| "eval_runtime": 186.0546, | |
| "eval_samples_per_second": 37.742, | |
| "eval_steps_per_second": 0.591, | |
| "eval_wer": 1.0, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 1.2825445684237526, | |
| "step": 20000, | |
| "total_flos": 2.2824984432894013e+19, | |
| "train_loss": 0.38660173568725587, | |
| "train_runtime": 15166.2226, | |
| "train_samples_per_second": 10.55, | |
| "train_steps_per_second": 1.319 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 20000, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 2, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 2.2824984432894013e+19, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |