| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 8.975317875841435, | |
| "global_step": 12000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "learning_rate": 2.0000000000000002e-07, | |
| "loss": 18.6192, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "learning_rate": 9.94e-05, | |
| "loss": 7.5397, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "eval_cer": 0.9810978889654326, | |
| "eval_clean_cer": 0.9774387393628106, | |
| "eval_loss": 3.571558952331543, | |
| "eval_runtime": 221.4433, | |
| "eval_samples_per_second": 24.679, | |
| "eval_steps_per_second": 1.544, | |
| "eval_wer": 1.0, | |
| "learning_rate": 9.94e-05, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "learning_rate": 9.613830613830615e-05, | |
| "loss": 1.7478, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "eval_cer": 0.22960449951399964, | |
| "eval_clean_cer": 0.17457099250398456, | |
| "eval_loss": 0.7702281475067139, | |
| "eval_runtime": 224.0929, | |
| "eval_samples_per_second": 24.387, | |
| "eval_steps_per_second": 1.526, | |
| "eval_wer": 0.8097043024248205, | |
| "learning_rate": 9.613830613830615e-05, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.12, | |
| "learning_rate": 9.225330225330226e-05, | |
| "loss": 0.7687, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.12, | |
| "eval_cer": 0.14148744420432777, | |
| "eval_clean_cer": 0.11820726398251868, | |
| "eval_loss": 0.4996742308139801, | |
| "eval_runtime": 221.7353, | |
| "eval_samples_per_second": 24.647, | |
| "eval_steps_per_second": 1.542, | |
| "eval_wer": 0.5391836939605402, | |
| "learning_rate": 9.225330225330226e-05, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.5, | |
| "learning_rate": 8.836829836829838e-05, | |
| "loss": 0.6064, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.5, | |
| "eval_cer": 0.12378853794105713, | |
| "eval_clean_cer": 0.10011770602778763, | |
| "eval_loss": 0.4269550144672394, | |
| "eval_runtime": 224.0891, | |
| "eval_samples_per_second": 24.388, | |
| "eval_steps_per_second": 1.526, | |
| "eval_wer": 0.49563231085897885, | |
| "learning_rate": 8.836829836829838e-05, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.87, | |
| "learning_rate": 8.448329448329449e-05, | |
| "loss": 0.5473, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.87, | |
| "eval_cer": 0.1105108194060415, | |
| "eval_clean_cer": 0.0898226524968884, | |
| "eval_loss": 0.38087567687034607, | |
| "eval_runtime": 223.664, | |
| "eval_samples_per_second": 24.434, | |
| "eval_steps_per_second": 1.529, | |
| "eval_wer": 0.4489432200411667, | |
| "learning_rate": 8.448329448329449e-05, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 2.24, | |
| "learning_rate": 8.05982905982906e-05, | |
| "loss": 0.454, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 2.24, | |
| "eval_cer": 0.10214972585805014, | |
| "eval_clean_cer": 0.08129037344912453, | |
| "eval_loss": 0.3585045337677002, | |
| "eval_runtime": 225.448, | |
| "eval_samples_per_second": 24.241, | |
| "eval_steps_per_second": 1.517, | |
| "eval_wer": 0.4256237762939907, | |
| "learning_rate": 8.05982905982906e-05, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 2.62, | |
| "learning_rate": 7.671328671328672e-05, | |
| "loss": 0.4219, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 2.62, | |
| "eval_cer": 0.0973510621231138, | |
| "eval_clean_cer": 0.07773103327870422, | |
| "eval_loss": 0.3375051021575928, | |
| "eval_runtime": 226.653, | |
| "eval_samples_per_second": 24.112, | |
| "eval_steps_per_second": 1.509, | |
| "eval_wer": 0.40634570008534565, | |
| "learning_rate": 7.671328671328672e-05, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 2.99, | |
| "learning_rate": 7.282828282828284e-05, | |
| "loss": 0.4075, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 2.99, | |
| "eval_cer": 0.09484084667868299, | |
| "eval_clean_cer": 0.07456029195600385, | |
| "eval_loss": 0.3274226784706116, | |
| "eval_runtime": 225.9771, | |
| "eval_samples_per_second": 24.184, | |
| "eval_steps_per_second": 1.513, | |
| "eval_wer": 0.4035845172950449, | |
| "learning_rate": 7.282828282828284e-05, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 3.37, | |
| "learning_rate": 6.894327894327895e-05, | |
| "loss": 0.3355, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 3.37, | |
| "eval_cer": 0.08976851283890268, | |
| "eval_clean_cer": 0.07285946801380934, | |
| "eval_loss": 0.3256622850894928, | |
| "eval_runtime": 227.1344, | |
| "eval_samples_per_second": 24.061, | |
| "eval_steps_per_second": 1.506, | |
| "eval_wer": 0.3781816356242783, | |
| "learning_rate": 6.894327894327895e-05, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 3.74, | |
| "learning_rate": 6.506604506604507e-05, | |
| "loss": 0.3203, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 3.74, | |
| "eval_cer": 0.08301169231930695, | |
| "eval_clean_cer": 0.06594916676522435, | |
| "eval_loss": 0.30244797468185425, | |
| "eval_runtime": 226.4786, | |
| "eval_samples_per_second": 24.13, | |
| "eval_steps_per_second": 1.51, | |
| "eval_wer": 0.35606707164014256, | |
| "learning_rate": 6.506604506604507e-05, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 4.11, | |
| "learning_rate": 6.118104118104118e-05, | |
| "loss": 0.3151, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 4.11, | |
| "eval_cer": 0.08304472146989157, | |
| "eval_clean_cer": 0.06525081521280011, | |
| "eval_loss": 0.3038225769996643, | |
| "eval_runtime": 228.8025, | |
| "eval_samples_per_second": 23.885, | |
| "eval_steps_per_second": 1.495, | |
| "eval_wer": 0.36061047241327376, | |
| "learning_rate": 6.118104118104118e-05, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 4.49, | |
| "learning_rate": 5.72960372960373e-05, | |
| "loss": 0.2713, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 4.49, | |
| "eval_cer": 0.08317211962214652, | |
| "eval_clean_cer": 0.06550424924392181, | |
| "eval_loss": 0.30516648292541504, | |
| "eval_runtime": 228.6779, | |
| "eval_samples_per_second": 23.898, | |
| "eval_steps_per_second": 1.496, | |
| "eval_wer": 0.3594557959736935, | |
| "learning_rate": 5.72960372960373e-05, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 4.86, | |
| "learning_rate": 5.3411033411033415e-05, | |
| "loss": 0.2685, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 4.86, | |
| "eval_cer": 0.07956250530825634, | |
| "eval_clean_cer": 0.06275026610573267, | |
| "eval_loss": 0.29325252771377563, | |
| "eval_runtime": 228.92, | |
| "eval_samples_per_second": 23.873, | |
| "eval_steps_per_second": 1.494, | |
| "eval_wer": 0.34364174908378936, | |
| "learning_rate": 5.3411033411033415e-05, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 5.24, | |
| "learning_rate": 4.952602952602953e-05, | |
| "loss": 0.2379, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 5.24, | |
| "eval_cer": 0.07627846405012881, | |
| "eval_clean_cer": 0.060807271867132986, | |
| "eval_loss": 0.30203360319137573, | |
| "eval_runtime": 229.4624, | |
| "eval_samples_per_second": 23.817, | |
| "eval_steps_per_second": 1.49, | |
| "eval_wer": 0.3361614538882474, | |
| "learning_rate": 4.952602952602953e-05, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 5.61, | |
| "learning_rate": 4.564879564879565e-05, | |
| "loss": 0.224, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 5.61, | |
| "eval_cer": 0.07449960836864307, | |
| "eval_clean_cer": 0.058937491904190675, | |
| "eval_loss": 0.2874354124069214, | |
| "eval_runtime": 229.9772, | |
| "eval_samples_per_second": 23.763, | |
| "eval_steps_per_second": 1.487, | |
| "eval_wer": 0.3264973141221949, | |
| "learning_rate": 4.564879564879565e-05, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 5.98, | |
| "learning_rate": 4.1763791763791764e-05, | |
| "loss": 0.2204, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 5.98, | |
| "eval_cer": 0.07241877188181227, | |
| "eval_clean_cer": 0.05759147560556654, | |
| "eval_loss": 0.29218629002571106, | |
| "eval_runtime": 229.2706, | |
| "eval_samples_per_second": 23.836, | |
| "eval_steps_per_second": 1.492, | |
| "eval_wer": 0.31909232391184295, | |
| "learning_rate": 4.1763791763791764e-05, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 6.36, | |
| "learning_rate": 3.787878787878788e-05, | |
| "loss": 0.1927, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 6.36, | |
| "eval_cer": 0.0719469268734606, | |
| "eval_clean_cer": 0.05684243724691796, | |
| "eval_loss": 0.3106653392314911, | |
| "eval_runtime": 228.9966, | |
| "eval_samples_per_second": 23.865, | |
| "eval_steps_per_second": 1.493, | |
| "eval_wer": 0.3163060394598122, | |
| "learning_rate": 3.787878787878788e-05, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 6.73, | |
| "learning_rate": 3.3993783993783994e-05, | |
| "loss": 0.1875, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 6.73, | |
| "eval_cer": 0.07033793539498145, | |
| "eval_clean_cer": 0.05535562426433733, | |
| "eval_loss": 0.30340105295181274, | |
| "eval_runtime": 229.2234, | |
| "eval_samples_per_second": 23.841, | |
| "eval_steps_per_second": 1.492, | |
| "eval_wer": 0.30837391435313016, | |
| "learning_rate": 3.3993783993783994e-05, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 7.11, | |
| "learning_rate": 3.010878010878011e-05, | |
| "loss": 0.1786, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 7.11, | |
| "eval_cer": 0.0701822265422254, | |
| "eval_clean_cer": 0.05527677812132169, | |
| "eval_loss": 0.3209647834300995, | |
| "eval_runtime": 230.4875, | |
| "eval_samples_per_second": 23.711, | |
| "eval_steps_per_second": 1.484, | |
| "eval_wer": 0.31070836889402076, | |
| "learning_rate": 3.010878010878011e-05, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 7.48, | |
| "learning_rate": 2.6231546231546235e-05, | |
| "loss": 0.1606, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 7.48, | |
| "eval_cer": 0.06884690516859023, | |
| "eval_clean_cer": 0.05414477278231143, | |
| "eval_loss": 0.3231320083141327, | |
| "eval_runtime": 228.6307, | |
| "eval_samples_per_second": 23.903, | |
| "eval_steps_per_second": 1.496, | |
| "eval_wer": 0.3062402731060796, | |
| "learning_rate": 2.6231546231546235e-05, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 7.85, | |
| "learning_rate": 2.234654234654235e-05, | |
| "loss": 0.1594, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 7.85, | |
| "eval_cer": 0.06802589485405834, | |
| "eval_clean_cer": 0.05349710803611153, | |
| "eval_loss": 0.3233787417411804, | |
| "eval_runtime": 229.7364, | |
| "eval_samples_per_second": 23.788, | |
| "eval_steps_per_second": 1.489, | |
| "eval_wer": 0.30325317536020885, | |
| "learning_rate": 2.234654234654235e-05, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 8.23, | |
| "learning_rate": 1.8461538461538465e-05, | |
| "loss": 0.1498, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 8.23, | |
| "eval_cer": 0.06797871035322317, | |
| "eval_clean_cer": 0.05303529491273422, | |
| "eval_loss": 0.3275805115699768, | |
| "eval_runtime": 229.9477, | |
| "eval_samples_per_second": 23.766, | |
| "eval_steps_per_second": 1.487, | |
| "eval_wer": 0.3035292936392389, | |
| "learning_rate": 1.8461538461538465e-05, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 8.6, | |
| "learning_rate": 1.4576534576534578e-05, | |
| "loss": 0.1396, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 8.6, | |
| "eval_cer": 0.06678022403200996, | |
| "eval_clean_cer": 0.05198776758409786, | |
| "eval_loss": 0.32653090357780457, | |
| "eval_runtime": 230.5216, | |
| "eval_samples_per_second": 23.707, | |
| "eval_steps_per_second": 1.484, | |
| "eval_wer": 0.2974797931623073, | |
| "learning_rate": 1.4576534576534578e-05, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 8.98, | |
| "learning_rate": 1.0691530691530691e-05, | |
| "loss": 0.142, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 8.98, | |
| "eval_cer": 0.06585540781564071, | |
| "eval_clean_cer": 0.05150905885864576, | |
| "eval_loss": 0.3235620856285095, | |
| "eval_runtime": 230.1535, | |
| "eval_samples_per_second": 23.745, | |
| "eval_steps_per_second": 1.486, | |
| "eval_wer": 0.2930367990360962, | |
| "learning_rate": 1.0691530691530691e-05, | |
| "step": 12000 | |
| } | |
| ], | |
| "max_steps": 13370, | |
| "num_train_epochs": 10, | |
| "total_flos": 5.065598164318501e+19, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |