| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 14.147606696533837, | |
| "eval_steps": 2000, | |
| "global_step": 30000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.24, | |
| "learning_rate": 0.00014879999999999998, | |
| "loss": 14.6823, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "learning_rate": 0.0002985, | |
| "loss": 4.0589, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "learning_rate": 0.0002951785714285714, | |
| "loss": 1.959, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "learning_rate": 0.00029030844155844153, | |
| "loss": 1.5961, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "eval_cer": 0.27122803037904003, | |
| "eval_loss": 1.068536400794983, | |
| "eval_runtime": 2688.7775, | |
| "eval_samples_per_second": 6.291, | |
| "eval_steps_per_second": 0.393, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.18, | |
| "learning_rate": 0.00028543831168831166, | |
| "loss": 1.4159, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.41, | |
| "learning_rate": 0.0002805681818181818, | |
| "loss": 1.3166, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.65, | |
| "learning_rate": 0.0002757077922077922, | |
| "loss": 1.2577, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 1.89, | |
| "learning_rate": 0.0002708376623376623, | |
| "loss": 1.2068, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.89, | |
| "eval_cer": 0.21168921247442604, | |
| "eval_loss": 0.8059473037719727, | |
| "eval_runtime": 2690.7607, | |
| "eval_samples_per_second": 6.286, | |
| "eval_steps_per_second": 0.393, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 2.12, | |
| "learning_rate": 0.00026596753246753245, | |
| "loss": 1.1463, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 2.36, | |
| "learning_rate": 0.0002610974025974026, | |
| "loss": 1.0865, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 2.59, | |
| "learning_rate": 0.000256237012987013, | |
| "loss": 1.0765, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 2.83, | |
| "learning_rate": 0.0002513668831168831, | |
| "loss": 1.0519, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 2.83, | |
| "eval_cer": 0.17837652734942439, | |
| "eval_loss": 0.6763827800750732, | |
| "eval_runtime": 2755.5593, | |
| "eval_samples_per_second": 6.138, | |
| "eval_steps_per_second": 0.384, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 3.07, | |
| "learning_rate": 0.00024649675324675324, | |
| "loss": 1.0204, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 3.3, | |
| "learning_rate": 0.00024162662337662335, | |
| "loss": 0.9806, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 3.54, | |
| "learning_rate": 0.0002367564935064935, | |
| "loss": 0.9693, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 3.77, | |
| "learning_rate": 0.00023189610389610387, | |
| "loss": 0.9522, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 3.77, | |
| "eval_cer": 0.1666790265096821, | |
| "eval_loss": 0.6237545609474182, | |
| "eval_runtime": 2657.2446, | |
| "eval_samples_per_second": 6.365, | |
| "eval_steps_per_second": 0.398, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 4.01, | |
| "learning_rate": 0.000227025974025974, | |
| "loss": 0.9543, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 4.24, | |
| "learning_rate": 0.00022215584415584413, | |
| "loss": 0.9079, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 4.48, | |
| "learning_rate": 0.00021728571428571427, | |
| "loss": 0.9006, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 4.72, | |
| "learning_rate": 0.00021242532467532466, | |
| "loss": 0.8855, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 4.72, | |
| "eval_cer": 0.15723901464423282, | |
| "eval_loss": 0.5901318788528442, | |
| "eval_runtime": 2755.8864, | |
| "eval_samples_per_second": 6.137, | |
| "eval_steps_per_second": 0.384, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 4.95, | |
| "learning_rate": 0.0002075551948051948, | |
| "loss": 0.8858, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 5.19, | |
| "learning_rate": 0.00020268506493506492, | |
| "loss": 0.8449, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 5.42, | |
| "learning_rate": 0.0001978246753246753, | |
| "loss": 0.8309, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 5.66, | |
| "learning_rate": 0.00019295454545454545, | |
| "loss": 0.8353, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 5.66, | |
| "eval_cer": 0.14726462133076249, | |
| "eval_loss": 0.5560417175292969, | |
| "eval_runtime": 2734.9608, | |
| "eval_samples_per_second": 6.184, | |
| "eval_steps_per_second": 0.387, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 5.89, | |
| "learning_rate": 0.00018808441558441555, | |
| "loss": 0.8255, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 6.13, | |
| "learning_rate": 0.0001832142857142857, | |
| "loss": 0.808, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 6.37, | |
| "learning_rate": 0.00017834415584415584, | |
| "loss": 0.7827, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 6.6, | |
| "learning_rate": 0.00017347402597402595, | |
| "loss": 0.7765, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 6.6, | |
| "eval_cer": 0.14182047400725015, | |
| "eval_loss": 0.5313007831573486, | |
| "eval_runtime": 2764.1566, | |
| "eval_samples_per_second": 6.119, | |
| "eval_steps_per_second": 0.383, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 6.84, | |
| "learning_rate": 0.00016861363636363634, | |
| "loss": 0.7766, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 7.07, | |
| "learning_rate": 0.00016375324675324673, | |
| "loss": 0.7511, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 7.31, | |
| "learning_rate": 0.0001588831168831169, | |
| "loss": 0.7287, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 7.55, | |
| "learning_rate": 0.000154012987012987, | |
| "loss": 0.7333, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 7.55, | |
| "eval_cer": 0.13388763594009692, | |
| "eval_loss": 0.5099530220031738, | |
| "eval_runtime": 2760.9388, | |
| "eval_samples_per_second": 6.126, | |
| "eval_steps_per_second": 0.383, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 7.78, | |
| "learning_rate": 0.00014914285714285713, | |
| "loss": 0.7285, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 8.02, | |
| "learning_rate": 0.00014427272727272726, | |
| "loss": 0.719, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 8.25, | |
| "learning_rate": 0.0001394025974025974, | |
| "loss": 0.6828, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 8.49, | |
| "learning_rate": 0.00013453246753246753, | |
| "loss": 0.6887, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 8.49, | |
| "eval_cer": 0.13040652250727414, | |
| "eval_loss": 0.49017849564552307, | |
| "eval_runtime": 2752.9478, | |
| "eval_samples_per_second": 6.144, | |
| "eval_steps_per_second": 0.384, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 8.72, | |
| "learning_rate": 0.00012966233766233766, | |
| "loss": 0.6893, | |
| "step": 18500 | |
| }, | |
| { | |
| "epoch": 8.96, | |
| "learning_rate": 0.00012480194805194805, | |
| "loss": 0.6834, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 9.2, | |
| "learning_rate": 0.00011993181818181817, | |
| "loss": 0.6572, | |
| "step": 19500 | |
| }, | |
| { | |
| "epoch": 9.43, | |
| "learning_rate": 0.00011506168831168829, | |
| "loss": 0.6547, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 9.43, | |
| "eval_cer": 0.12521102614324914, | |
| "eval_loss": 0.47851961851119995, | |
| "eval_runtime": 2750.0743, | |
| "eval_samples_per_second": 6.15, | |
| "eval_steps_per_second": 0.385, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 9.67, | |
| "learning_rate": 0.00011019155844155843, | |
| "loss": 0.6438, | |
| "step": 20500 | |
| }, | |
| { | |
| "epoch": 9.9, | |
| "learning_rate": 0.00010533116883116881, | |
| "loss": 0.6419, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 10.14, | |
| "learning_rate": 0.00010046103896103896, | |
| "loss": 0.6238, | |
| "step": 21500 | |
| }, | |
| { | |
| "epoch": 10.37, | |
| "learning_rate": 9.560064935064933e-05, | |
| "loss": 0.612, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 10.37, | |
| "eval_cer": 0.1199806313989452, | |
| "eval_loss": 0.45936959981918335, | |
| "eval_runtime": 2703.4007, | |
| "eval_samples_per_second": 6.257, | |
| "eval_steps_per_second": 0.391, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 10.61, | |
| "learning_rate": 9.073051948051948e-05, | |
| "loss": 0.6138, | |
| "step": 22500 | |
| }, | |
| { | |
| "epoch": 10.85, | |
| "learning_rate": 8.58603896103896e-05, | |
| "loss": 0.6128, | |
| "step": 23000 | |
| }, | |
| { | |
| "epoch": 11.08, | |
| "learning_rate": 8.099025974025973e-05, | |
| "loss": 0.6003, | |
| "step": 23500 | |
| }, | |
| { | |
| "epoch": 11.32, | |
| "learning_rate": 7.612987012987012e-05, | |
| "loss": 0.5855, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 11.32, | |
| "eval_cer": 0.1176489833665595, | |
| "eval_loss": 0.4468931555747986, | |
| "eval_runtime": 2737.3716, | |
| "eval_samples_per_second": 6.179, | |
| "eval_steps_per_second": 0.387, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 11.55, | |
| "learning_rate": 7.125974025974026e-05, | |
| "loss": 0.5766, | |
| "step": 24500 | |
| }, | |
| { | |
| "epoch": 11.79, | |
| "learning_rate": 6.638961038961039e-05, | |
| "loss": 0.5821, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 12.03, | |
| "learning_rate": 6.151948051948051e-05, | |
| "loss": 0.5765, | |
| "step": 25500 | |
| }, | |
| { | |
| "epoch": 12.26, | |
| "learning_rate": 5.664935064935064e-05, | |
| "loss": 0.5538, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 12.26, | |
| "eval_cer": 0.11556380514489371, | |
| "eval_loss": 0.4398212134838104, | |
| "eval_runtime": 2705.0289, | |
| "eval_samples_per_second": 6.253, | |
| "eval_steps_per_second": 0.391, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 12.5, | |
| "learning_rate": 5.178896103896104e-05, | |
| "loss": 0.5612, | |
| "step": 26500 | |
| }, | |
| { | |
| "epoch": 12.73, | |
| "learning_rate": 4.691883116883116e-05, | |
| "loss": 0.5511, | |
| "step": 27000 | |
| }, | |
| { | |
| "epoch": 12.97, | |
| "learning_rate": 4.2048701298701296e-05, | |
| "loss": 0.5497, | |
| "step": 27500 | |
| }, | |
| { | |
| "epoch": 13.2, | |
| "learning_rate": 3.717857142857143e-05, | |
| "loss": 0.5341, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 13.2, | |
| "eval_cer": 0.11241640747348813, | |
| "eval_loss": 0.43176981806755066, | |
| "eval_runtime": 2707.434, | |
| "eval_samples_per_second": 6.247, | |
| "eval_steps_per_second": 0.391, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 13.44, | |
| "learning_rate": 3.231818181818181e-05, | |
| "loss": 0.5317, | |
| "step": 28500 | |
| }, | |
| { | |
| "epoch": 13.68, | |
| "learning_rate": 2.7448051948051945e-05, | |
| "loss": 0.5315, | |
| "step": 29000 | |
| }, | |
| { | |
| "epoch": 13.91, | |
| "learning_rate": 2.2577922077922077e-05, | |
| "loss": 0.5265, | |
| "step": 29500 | |
| }, | |
| { | |
| "epoch": 14.15, | |
| "learning_rate": 1.7707792207792207e-05, | |
| "loss": 0.5229, | |
| "step": 30000 | |
| }, | |
| { | |
| "epoch": 14.15, | |
| "eval_cer": 0.11112952970070276, | |
| "eval_loss": 0.4251420199871063, | |
| "eval_runtime": 2738.5837, | |
| "eval_samples_per_second": 6.176, | |
| "eval_steps_per_second": 0.386, | |
| "step": 30000 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 31800, | |
| "num_train_epochs": 15, | |
| "save_steps": 2000, | |
| "total_flos": 2.014685631404017e+20, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |