| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 39.95179987797437, | |
| "eval_steps": 100.0, | |
| "global_step": 32760, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.6101281269066504, | |
| "grad_norm": 50.660884857177734, | |
| "learning_rate": 1.188e-06, | |
| "loss": 40.3051, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_cer": 1.1284080132764343, | |
| "eval_loss": 14.975272178649902, | |
| "eval_runtime": 87.9488, | |
| "eval_samples_per_second": 77.773, | |
| "eval_steps_per_second": 9.722, | |
| "eval_wer": 1.0, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 1.2196461256863942, | |
| "grad_norm": 66.72471618652344, | |
| "learning_rate": 2.3880000000000003e-06, | |
| "loss": 15.5316, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.8297742525930445, | |
| "grad_norm": 58.356754302978516, | |
| "learning_rate": 3.588e-06, | |
| "loss": 12.3062, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_cer": 1.1284376481744902, | |
| "eval_loss": 9.953412055969238, | |
| "eval_runtime": 86.9725, | |
| "eval_samples_per_second": 78.646, | |
| "eval_steps_per_second": 9.831, | |
| "eval_wer": 1.0, | |
| "step": 1640 | |
| }, | |
| { | |
| "epoch": 2.4392922513727884, | |
| "grad_norm": 37.668678283691406, | |
| "learning_rate": 4.788e-06, | |
| "loss": 9.0113, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_cer": 1.1283783783783783, | |
| "eval_loss": 4.837707996368408, | |
| "eval_runtime": 90.6105, | |
| "eval_samples_per_second": 75.488, | |
| "eval_steps_per_second": 9.436, | |
| "eval_wer": 1.0, | |
| "step": 2460 | |
| }, | |
| { | |
| "epoch": 3.048810250152532, | |
| "grad_norm": 10.56650161743164, | |
| "learning_rate": 5.988e-06, | |
| "loss": 5.8841, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 3.6589383770591826, | |
| "grad_norm": 1.526659369468689, | |
| "learning_rate": 7.1880000000000005e-06, | |
| "loss": 3.9478, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_cer": 1.1284080132764343, | |
| "eval_loss": 3.201613426208496, | |
| "eval_runtime": 92.4896, | |
| "eval_samples_per_second": 73.954, | |
| "eval_steps_per_second": 9.244, | |
| "eval_wer": 1.0, | |
| "step": 3280 | |
| }, | |
| { | |
| "epoch": 4.268456375838926, | |
| "grad_norm": 1.9021031856536865, | |
| "learning_rate": 8.388e-06, | |
| "loss": 3.2121, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 4.878584502745577, | |
| "grad_norm": 2.086733102798462, | |
| "learning_rate": 9.588e-06, | |
| "loss": 2.7396, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_cer": 1.1284376481744902, | |
| "eval_loss": 2.4831321239471436, | |
| "eval_runtime": 89.7632, | |
| "eval_samples_per_second": 76.201, | |
| "eval_steps_per_second": 9.525, | |
| "eval_wer": 1.0, | |
| "step": 4100 | |
| }, | |
| { | |
| "epoch": 5.48810250152532, | |
| "grad_norm": 4.460277557373047, | |
| "learning_rate": 1.0787999999999999e-05, | |
| "loss": 2.2509, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_cer": 1.1053816974869606, | |
| "eval_loss": 0.9197890758514404, | |
| "eval_runtime": 86.71, | |
| "eval_samples_per_second": 78.884, | |
| "eval_steps_per_second": 9.86, | |
| "eval_wer": 0.9998538011695907, | |
| "step": 4920 | |
| }, | |
| { | |
| "epoch": 6.097620500305064, | |
| "grad_norm": 3.249329090118408, | |
| "learning_rate": 1.1988000000000001e-05, | |
| "loss": 1.2958, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 6.707748627211714, | |
| "grad_norm": 3.6012656688690186, | |
| "learning_rate": 1.3188e-05, | |
| "loss": 0.6886, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_cer": 1.0272641062114747, | |
| "eval_loss": 0.3222425580024719, | |
| "eval_runtime": 89.592, | |
| "eval_samples_per_second": 76.346, | |
| "eval_steps_per_second": 9.543, | |
| "eval_wer": 0.9991228070175439, | |
| "step": 5740 | |
| }, | |
| { | |
| "epoch": 7.317266625991458, | |
| "grad_norm": 3.2955262660980225, | |
| "learning_rate": 1.4388000000000002e-05, | |
| "loss": 0.439, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 7.927394752898109, | |
| "grad_norm": 25.04276466369629, | |
| "learning_rate": 1.5588e-05, | |
| "loss": 0.318, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_cer": 1.0191441441441442, | |
| "eval_loss": 0.16009780764579773, | |
| "eval_runtime": 88.8354, | |
| "eval_samples_per_second": 76.996, | |
| "eval_steps_per_second": 9.625, | |
| "eval_wer": 0.9991228070175439, | |
| "step": 6560 | |
| }, | |
| { | |
| "epoch": 8.536912751677852, | |
| "grad_norm": 2.714284896850586, | |
| "learning_rate": 1.6788e-05, | |
| "loss": 0.255, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_cer": 1.0176327643432908, | |
| "eval_loss": 0.09021047502756119, | |
| "eval_runtime": 89.3568, | |
| "eval_samples_per_second": 76.547, | |
| "eval_steps_per_second": 9.568, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 7380 | |
| }, | |
| { | |
| "epoch": 9.146430750457595, | |
| "grad_norm": 4.7839813232421875, | |
| "learning_rate": 1.7988e-05, | |
| "loss": 0.2207, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 9.756558877364247, | |
| "grad_norm": 2.4737026691436768, | |
| "learning_rate": 1.9188e-05, | |
| "loss": 0.1835, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_cer": 1.019736842105263, | |
| "eval_loss": 0.07156170159578323, | |
| "eval_runtime": 88.4936, | |
| "eval_samples_per_second": 77.294, | |
| "eval_steps_per_second": 9.662, | |
| "eval_wer": 0.9991228070175439, | |
| "step": 8200 | |
| }, | |
| { | |
| "epoch": 10.36607687614399, | |
| "grad_norm": 6.031515121459961, | |
| "learning_rate": 2.0388e-05, | |
| "loss": 0.1691, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 10.97620500305064, | |
| "grad_norm": 3.6490681171417236, | |
| "learning_rate": 2.1588e-05, | |
| "loss": 0.16, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "eval_cer": 1.017514224751067, | |
| "eval_loss": 0.04918373003602028, | |
| "eval_runtime": 88.5165, | |
| "eval_samples_per_second": 77.274, | |
| "eval_steps_per_second": 9.659, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 9020 | |
| }, | |
| { | |
| "epoch": 11.585723001830385, | |
| "grad_norm": 4.373748779296875, | |
| "learning_rate": 2.2788000000000003e-05, | |
| "loss": 0.1444, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_cer": 1.016151019440493, | |
| "eval_loss": 0.03436645492911339, | |
| "eval_runtime": 88.2172, | |
| "eval_samples_per_second": 77.536, | |
| "eval_steps_per_second": 9.692, | |
| "eval_wer": 0.9989766081871345, | |
| "step": 9840 | |
| }, | |
| { | |
| "epoch": 12.195241000610128, | |
| "grad_norm": 0.14861944317817688, | |
| "learning_rate": 2.3988e-05, | |
| "loss": 0.1382, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 12.805369127516778, | |
| "grad_norm": 4.033771991729736, | |
| "learning_rate": 2.5188e-05, | |
| "loss": 0.1307, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "eval_cer": 1.0176327643432908, | |
| "eval_loss": 0.038768209517002106, | |
| "eval_runtime": 95.1356, | |
| "eval_samples_per_second": 71.897, | |
| "eval_steps_per_second": 8.987, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 10660 | |
| }, | |
| { | |
| "epoch": 13.414887126296522, | |
| "grad_norm": 2.4692184925079346, | |
| "learning_rate": 2.63856e-05, | |
| "loss": 0.1346, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_cer": 1.0166548127074442, | |
| "eval_loss": 0.0276629775762558, | |
| "eval_runtime": 87.0897, | |
| "eval_samples_per_second": 78.54, | |
| "eval_steps_per_second": 9.817, | |
| "eval_wer": 0.9989766081871345, | |
| "step": 11480 | |
| }, | |
| { | |
| "epoch": 14.024405125076266, | |
| "grad_norm": 3.765444278717041, | |
| "learning_rate": 2.7585600000000002e-05, | |
| "loss": 0.1287, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 14.634533251982916, | |
| "grad_norm": 1.3012793064117432, | |
| "learning_rate": 2.87856e-05, | |
| "loss": 0.1257, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "eval_cer": 1.0164473684210527, | |
| "eval_loss": 0.02670404687523842, | |
| "eval_runtime": 92.4139, | |
| "eval_samples_per_second": 74.015, | |
| "eval_steps_per_second": 9.252, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 12300 | |
| }, | |
| { | |
| "epoch": 15.24405125076266, | |
| "grad_norm": 6.9578728675842285, | |
| "learning_rate": 2.99856e-05, | |
| "loss": 0.1276, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 15.854179377669311, | |
| "grad_norm": 3.3747990131378174, | |
| "learning_rate": 2.9956013070043084e-05, | |
| "loss": 0.1157, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "eval_cer": 1.0165955429113325, | |
| "eval_loss": 0.02758130058646202, | |
| "eval_runtime": 89.1614, | |
| "eval_samples_per_second": 76.715, | |
| "eval_steps_per_second": 9.589, | |
| "eval_wer": 0.9989766081871345, | |
| "step": 13120 | |
| }, | |
| { | |
| "epoch": 16.463697376449055, | |
| "grad_norm": 2.6571855545043945, | |
| "learning_rate": 2.9822174136311704e-05, | |
| "loss": 0.1104, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 17.0, | |
| "eval_cer": 1.0171586059743956, | |
| "eval_loss": 0.03412469103932381, | |
| "eval_runtime": 88.3916, | |
| "eval_samples_per_second": 77.383, | |
| "eval_steps_per_second": 9.673, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 13940 | |
| }, | |
| { | |
| "epoch": 17.0732153752288, | |
| "grad_norm": 0.15126635134220123, | |
| "learning_rate": 2.9599814696946643e-05, | |
| "loss": 0.1152, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 17.683343502135447, | |
| "grad_norm": 1.5255239009857178, | |
| "learning_rate": 2.9289379955813937e-05, | |
| "loss": 0.1086, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 18.0, | |
| "eval_cer": 1.016743717401612, | |
| "eval_loss": 0.023562652990221977, | |
| "eval_runtime": 99.8828, | |
| "eval_samples_per_second": 68.48, | |
| "eval_steps_per_second": 8.56, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 14760 | |
| }, | |
| { | |
| "epoch": 18.29286150091519, | |
| "grad_norm": 4.014769554138184, | |
| "learning_rate": 2.8893091974003682e-05, | |
| "loss": 0.1007, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 18.902989627821842, | |
| "grad_norm": 1.9409995079040527, | |
| "learning_rate": 2.84143727148899e-05, | |
| "loss": 0.1072, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 19.0, | |
| "eval_cer": 1.0175438596491229, | |
| "eval_loss": 0.030639806762337685, | |
| "eval_runtime": 84.6618, | |
| "eval_samples_per_second": 80.792, | |
| "eval_steps_per_second": 10.099, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 15580 | |
| }, | |
| { | |
| "epoch": 19.512507626601586, | |
| "grad_norm": 13.441529273986816, | |
| "learning_rate": 2.7855379321676933e-05, | |
| "loss": 0.128, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "eval_cer": 1.0160324798482694, | |
| "eval_loss": 0.024129284545779228, | |
| "eval_runtime": 90.1768, | |
| "eval_samples_per_second": 75.851, | |
| "eval_steps_per_second": 9.481, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 16400 | |
| }, | |
| { | |
| "epoch": 20.12202562538133, | |
| "grad_norm": 2.291337490081787, | |
| "learning_rate": 2.7218107759869366e-05, | |
| "loss": 0.1062, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 20.73215375228798, | |
| "grad_norm": 4.577757835388184, | |
| "learning_rate": 2.650742754426605e-05, | |
| "loss": 0.0987, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 21.0, | |
| "eval_cer": 1.0165066382171646, | |
| "eval_loss": 0.02422538958489895, | |
| "eval_runtime": 97.4522, | |
| "eval_samples_per_second": 70.188, | |
| "eval_steps_per_second": 8.774, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 17220 | |
| }, | |
| { | |
| "epoch": 21.341671751067725, | |
| "grad_norm": 6.208291530609131, | |
| "learning_rate": 2.5727608573195923e-05, | |
| "loss": 0.1035, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 21.951799877974373, | |
| "grad_norm": 9.813406944274902, | |
| "learning_rate": 2.4883336143432908e-05, | |
| "loss": 0.1031, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 22.0, | |
| "eval_cer": 1.0159435751541015, | |
| "eval_loss": 0.02291141264140606, | |
| "eval_runtime": 85.7962, | |
| "eval_samples_per_second": 79.724, | |
| "eval_steps_per_second": 9.965, | |
| "eval_wer": 0.9989766081871345, | |
| "step": 18040 | |
| }, | |
| { | |
| "epoch": 22.561317876754117, | |
| "grad_norm": 0.14426590502262115, | |
| "learning_rate": 2.3979682800065307e-05, | |
| "loss": 0.0903, | |
| "step": 18500 | |
| }, | |
| { | |
| "epoch": 23.0, | |
| "eval_cer": 1.016239924134661, | |
| "eval_loss": 0.02606978453695774, | |
| "eval_runtime": 103.6933, | |
| "eval_samples_per_second": 65.964, | |
| "eval_steps_per_second": 8.245, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 18860 | |
| }, | |
| { | |
| "epoch": 23.17083587553386, | |
| "grad_norm": 33.88288497924805, | |
| "learning_rate": 2.3024043139715204e-05, | |
| "loss": 0.0964, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 23.780964002440513, | |
| "grad_norm": 2.688063859939575, | |
| "learning_rate": 2.2018330526045242e-05, | |
| "loss": 0.0895, | |
| "step": 19500 | |
| }, | |
| { | |
| "epoch": 24.0, | |
| "eval_cer": 1.0176920341394025, | |
| "eval_loss": 0.03217785060405731, | |
| "eval_runtime": 89.5142, | |
| "eval_samples_per_second": 76.412, | |
| "eval_steps_per_second": 9.552, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 19680 | |
| }, | |
| { | |
| "epoch": 24.390482001220256, | |
| "grad_norm": 1.0760211944580078, | |
| "learning_rate": 2.0970450483020733e-05, | |
| "loss": 0.0844, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 25.0, | |
| "grad_norm": 0.8288090229034424, | |
| "learning_rate": 1.9886698867971603e-05, | |
| "loss": 0.0835, | |
| "step": 20500 | |
| }, | |
| { | |
| "epoch": 25.0, | |
| "eval_cer": 1.0152027027027026, | |
| "eval_loss": 0.018799621611833572, | |
| "eval_runtime": 88.0548, | |
| "eval_samples_per_second": 77.679, | |
| "eval_steps_per_second": 9.71, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 20500 | |
| }, | |
| { | |
| "epoch": 25.610128126906652, | |
| "grad_norm": 2.2765700817108154, | |
| "learning_rate": 1.877358706127469e-05, | |
| "loss": 0.0744, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 26.0, | |
| "eval_cer": 1.0149063537221432, | |
| "eval_loss": 0.01787102408707142, | |
| "eval_runtime": 95.8003, | |
| "eval_samples_per_second": 71.399, | |
| "eval_steps_per_second": 8.925, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 21320 | |
| }, | |
| { | |
| "epoch": 26.219646125686396, | |
| "grad_norm": 2.9212796688079834, | |
| "learning_rate": 1.7637802844774755e-05, | |
| "loss": 0.0738, | |
| "step": 21500 | |
| }, | |
| { | |
| "epoch": 26.829774252593044, | |
| "grad_norm": 3.6706600189208984, | |
| "learning_rate": 1.6486170220352805e-05, | |
| "loss": 0.0728, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 27.0, | |
| "eval_cer": 1.014521100047416, | |
| "eval_loss": 0.010682709515094757, | |
| "eval_runtime": 90.6475, | |
| "eval_samples_per_second": 75.457, | |
| "eval_steps_per_second": 9.432, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 22140 | |
| }, | |
| { | |
| "epoch": 27.439292251372787, | |
| "grad_norm": 1.4717997312545776, | |
| "learning_rate": 1.5325608410059234e-05, | |
| "loss": 0.0704, | |
| "step": 22500 | |
| }, | |
| { | |
| "epoch": 28.0, | |
| "eval_cer": 1.0149063537221432, | |
| "eval_loss": 0.016120394691824913, | |
| "eval_runtime": 91.4543, | |
| "eval_samples_per_second": 74.791, | |
| "eval_steps_per_second": 9.349, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 22960 | |
| }, | |
| { | |
| "epoch": 28.04881025015253, | |
| "grad_norm": 0.24731288850307465, | |
| "learning_rate": 1.4163090284146517e-05, | |
| "loss": 0.0651, | |
| "step": 23000 | |
| }, | |
| { | |
| "epoch": 28.658938377059183, | |
| "grad_norm": 1.4096115827560425, | |
| "learning_rate": 1.3007905796568247e-05, | |
| "loss": 0.068, | |
| "step": 23500 | |
| }, | |
| { | |
| "epoch": 29.0, | |
| "eval_cer": 1.0149656235182551, | |
| "eval_loss": 0.014005111530423164, | |
| "eval_runtime": 85.3622, | |
| "eval_samples_per_second": 80.129, | |
| "eval_steps_per_second": 10.016, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 23780 | |
| }, | |
| { | |
| "epoch": 29.268456375838927, | |
| "grad_norm": 1.2507057189941406, | |
| "learning_rate": 1.1862367835867989e-05, | |
| "loss": 0.062, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 29.878584502745575, | |
| "grad_norm": 0.42576301097869873, | |
| "learning_rate": 1.073568135019168e-05, | |
| "loss": 0.0635, | |
| "step": 24500 | |
| }, | |
| { | |
| "epoch": 30.0, | |
| "eval_cer": 1.0147878141299194, | |
| "eval_loss": 0.017921432852745056, | |
| "eval_runtime": 86.6261, | |
| "eval_samples_per_second": 78.96, | |
| "eval_steps_per_second": 9.87, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 24600 | |
| }, | |
| { | |
| "epoch": 30.48810250152532, | |
| "grad_norm": 1.0933780670166016, | |
| "learning_rate": 9.634615680568962e-06, | |
| "loss": 0.0606, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 31.0, | |
| "eval_cer": 1.0146100047415836, | |
| "eval_loss": 0.016976628452539444, | |
| "eval_runtime": 85.312, | |
| "eval_samples_per_second": 80.176, | |
| "eval_steps_per_second": 10.022, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 25420 | |
| }, | |
| { | |
| "epoch": 31.097620500305062, | |
| "grad_norm": 2.098126173019409, | |
| "learning_rate": 8.56578623342252e-06, | |
| "loss": 0.0588, | |
| "step": 25500 | |
| }, | |
| { | |
| "epoch": 31.707748627211714, | |
| "grad_norm": 0.9397180080413818, | |
| "learning_rate": 7.535614733981355e-06, | |
| "loss": 0.0549, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 32.0, | |
| "eval_cer": 1.0146100047415836, | |
| "eval_loss": 0.012712378054857254, | |
| "eval_runtime": 78.6951, | |
| "eval_samples_per_second": 86.918, | |
| "eval_steps_per_second": 10.865, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 26240 | |
| }, | |
| { | |
| "epoch": 32.31726662599146, | |
| "grad_norm": 3.3270745277404785, | |
| "learning_rate": 6.550290643366546e-06, | |
| "loss": 0.0545, | |
| "step": 26500 | |
| }, | |
| { | |
| "epoch": 32.92739475289811, | |
| "grad_norm": 1.39926016330719, | |
| "learning_rate": 5.615733971162722e-06, | |
| "loss": 0.0557, | |
| "step": 27000 | |
| }, | |
| { | |
| "epoch": 33.0, | |
| "eval_cer": 1.0145803698435278, | |
| "eval_loss": 0.011217311024665833, | |
| "eval_runtime": 76.5204, | |
| "eval_samples_per_second": 89.388, | |
| "eval_steps_per_second": 11.173, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 27060 | |
| }, | |
| { | |
| "epoch": 33.53691275167785, | |
| "grad_norm": 0.6240960955619812, | |
| "learning_rate": 4.737559706904321e-06, | |
| "loss": 0.0525, | |
| "step": 27500 | |
| }, | |
| { | |
| "epoch": 34.0, | |
| "eval_cer": 1.0145507349454719, | |
| "eval_loss": 0.013999322429299355, | |
| "eval_runtime": 76.0714, | |
| "eval_samples_per_second": 89.915, | |
| "eval_steps_per_second": 11.239, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 27880 | |
| }, | |
| { | |
| "epoch": 34.1464307504576, | |
| "grad_norm": 0.5480217933654785, | |
| "learning_rate": 3.924180691546633e-06, | |
| "loss": 0.0509, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 34.75655887736425, | |
| "grad_norm": 0.09108582139015198, | |
| "learning_rate": 3.1739539781329047e-06, | |
| "loss": 0.0478, | |
| "step": 28500 | |
| }, | |
| { | |
| "epoch": 35.0, | |
| "eval_cer": 1.0145507349454719, | |
| "eval_loss": 0.012541352771222591, | |
| "eval_runtime": 75.8184, | |
| "eval_samples_per_second": 90.216, | |
| "eval_steps_per_second": 11.277, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 28700 | |
| }, | |
| { | |
| "epoch": 35.36607687614399, | |
| "grad_norm": 5.189497470855713, | |
| "learning_rate": 2.4947803389966218e-06, | |
| "loss": 0.051, | |
| "step": 29000 | |
| }, | |
| { | |
| "epoch": 35.97620500305064, | |
| "grad_norm": 1.7753878831863403, | |
| "learning_rate": 1.8907403751213792e-06, | |
| "loss": 0.0475, | |
| "step": 29500 | |
| }, | |
| { | |
| "epoch": 36.0, | |
| "eval_cer": 1.0145803698435278, | |
| "eval_loss": 0.012446345761418343, | |
| "eval_runtime": 75.8714, | |
| "eval_samples_per_second": 90.153, | |
| "eval_steps_per_second": 11.269, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 29520 | |
| }, | |
| { | |
| "epoch": 36.58572300183038, | |
| "grad_norm": 1.131529688835144, | |
| "learning_rate": 1.3654632704576153e-06, | |
| "loss": 0.0455, | |
| "step": 30000 | |
| }, | |
| { | |
| "epoch": 37.0, | |
| "eval_cer": 1.0141951161688003, | |
| "eval_loss": 0.011377622373402119, | |
| "eval_runtime": 76.0985, | |
| "eval_samples_per_second": 89.884, | |
| "eval_steps_per_second": 11.235, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 30340 | |
| }, | |
| { | |
| "epoch": 37.195241000610125, | |
| "grad_norm": 0.012475825846195221, | |
| "learning_rate": 9.229080945215807e-07, | |
| "loss": 0.0502, | |
| "step": 30500 | |
| }, | |
| { | |
| "epoch": 37.80536912751678, | |
| "grad_norm": 0.05113031342625618, | |
| "learning_rate": 5.639609229205172e-07, | |
| "loss": 0.0444, | |
| "step": 31000 | |
| }, | |
| { | |
| "epoch": 38.0, | |
| "eval_cer": 1.0142247510668563, | |
| "eval_loss": 0.011828911490738392, | |
| "eval_runtime": 94.6786, | |
| "eval_samples_per_second": 72.244, | |
| "eval_steps_per_second": 9.031, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 31160 | |
| }, | |
| { | |
| "epoch": 38.41488712629652, | |
| "grad_norm": 2.288602113723755, | |
| "learning_rate": 2.9174814715380303e-07, | |
| "loss": 0.0476, | |
| "step": 31500 | |
| }, | |
| { | |
| "epoch": 39.0, | |
| "eval_cer": 1.0141951161688003, | |
| "eval_loss": 0.011857305653393269, | |
| "eval_runtime": 74.477, | |
| "eval_samples_per_second": 91.84, | |
| "eval_steps_per_second": 11.48, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 31980 | |
| }, | |
| { | |
| "epoch": 39.024405125076264, | |
| "grad_norm": 1.6855770349502563, | |
| "learning_rate": 1.0790527198271116e-07, | |
| "loss": 0.0451, | |
| "step": 32000 | |
| }, | |
| { | |
| "epoch": 39.634533251982916, | |
| "grad_norm": 2.309912919998169, | |
| "learning_rate": 1.3536859442666582e-08, | |
| "loss": 0.0464, | |
| "step": 32500 | |
| }, | |
| { | |
| "epoch": 39.95179987797437, | |
| "eval_cer": 1.0143432906590801, | |
| "eval_loss": 0.011460067704319954, | |
| "eval_runtime": 89.8642, | |
| "eval_samples_per_second": 76.115, | |
| "eval_steps_per_second": 9.514, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 32760 | |
| }, | |
| { | |
| "epoch": 39.95179987797437, | |
| "step": 32760, | |
| "total_flos": 2.3262081232451936e+19, | |
| "train_loss": 1.5701773451070355, | |
| "train_runtime": 35107.1781, | |
| "train_samples_per_second": 59.748, | |
| "train_steps_per_second": 0.933 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 32760, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 40, | |
| "save_steps": 400, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 2.3262081232451936e+19, | |
| "train_batch_size": 32, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |