| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 100.0, | |
| "global_step": 5700, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.75, | |
| "learning_rate": 5.88e-05, | |
| "loss": 8.0872, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 3.51, | |
| "learning_rate": 0.0001188, | |
| "loss": 3.3933, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 5.26, | |
| "learning_rate": 0.00017879999999999998, | |
| "loss": 3.1028, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 7.02, | |
| "learning_rate": 0.0002388, | |
| "loss": 2.635, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 8.77, | |
| "learning_rate": 0.0002988, | |
| "loss": 1.8032, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 8.77, | |
| "eval_loss": 0.8058779239654541, | |
| "eval_runtime": 34.1343, | |
| "eval_samples_per_second": 23.73, | |
| "eval_steps_per_second": 0.762, | |
| "eval_wer": 0.8351966075558982, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 10.53, | |
| "learning_rate": 0.0002943461538461538, | |
| "loss": 1.5861, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 12.28, | |
| "learning_rate": 0.00028857692307692303, | |
| "loss": 1.4828, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 14.04, | |
| "learning_rate": 0.0002828076923076923, | |
| "loss": 1.3835, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 15.79, | |
| "learning_rate": 0.0002770384615384615, | |
| "loss": 1.3352, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 17.54, | |
| "learning_rate": 0.00027126923076923075, | |
| "loss": 1.2608, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 17.54, | |
| "eval_loss": 0.5827550292015076, | |
| "eval_runtime": 34.2344, | |
| "eval_samples_per_second": 23.66, | |
| "eval_steps_per_second": 0.759, | |
| "eval_wer": 0.676946800308404, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 19.3, | |
| "learning_rate": 0.0002655, | |
| "loss": 1.239, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 21.05, | |
| "learning_rate": 0.0002597307692307692, | |
| "loss": 1.2078, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 22.81, | |
| "learning_rate": 0.00025396153846153846, | |
| "loss": 1.1695, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 24.56, | |
| "learning_rate": 0.00024819230769230767, | |
| "loss": 1.1397, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 26.32, | |
| "learning_rate": 0.0002424807692307692, | |
| "loss": 1.1337, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 26.32, | |
| "eval_loss": 0.6892158389091492, | |
| "eval_runtime": 34.4547, | |
| "eval_samples_per_second": 23.509, | |
| "eval_steps_per_second": 0.755, | |
| "eval_wer": 0.6908249807247494, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 28.07, | |
| "learning_rate": 0.00023671153846153845, | |
| "loss": 1.1085, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 29.82, | |
| "learning_rate": 0.00023094230769230768, | |
| "loss": 1.0829, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 31.58, | |
| "learning_rate": 0.00022517307692307688, | |
| "loss": 1.0601, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 33.33, | |
| "learning_rate": 0.00021940384615384614, | |
| "loss": 1.0398, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 35.09, | |
| "learning_rate": 0.00021363461538461537, | |
| "loss": 1.0457, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 35.09, | |
| "eval_loss": 0.7076693177223206, | |
| "eval_runtime": 34.0487, | |
| "eval_samples_per_second": 23.789, | |
| "eval_steps_per_second": 0.764, | |
| "eval_wer": 0.6781033153430994, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 36.84, | |
| "learning_rate": 0.0002078653846153846, | |
| "loss": 1.0369, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 38.6, | |
| "learning_rate": 0.00020209615384615383, | |
| "loss": 0.9969, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 40.35, | |
| "learning_rate": 0.00019632692307692306, | |
| "loss": 0.9822, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 42.11, | |
| "learning_rate": 0.0001905576923076923, | |
| "loss": 0.9744, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 43.86, | |
| "learning_rate": 0.00018478846153846152, | |
| "loss": 0.97, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 43.86, | |
| "eval_loss": 0.5993424654006958, | |
| "eval_runtime": 34.4764, | |
| "eval_samples_per_second": 23.494, | |
| "eval_steps_per_second": 0.754, | |
| "eval_wer": 0.6227833461835004, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 45.61, | |
| "learning_rate": 0.00017901923076923073, | |
| "loss": 0.9295, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 47.37, | |
| "learning_rate": 0.00017324999999999998, | |
| "loss": 0.9354, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 49.12, | |
| "learning_rate": 0.0001674807692307692, | |
| "loss": 0.9289, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 50.88, | |
| "learning_rate": 0.00016171153846153847, | |
| "loss": 0.8967, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 52.63, | |
| "learning_rate": 0.00015594230769230767, | |
| "loss": 0.8767, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 52.63, | |
| "eval_loss": 0.7213280200958252, | |
| "eval_runtime": 33.9189, | |
| "eval_samples_per_second": 23.88, | |
| "eval_steps_per_second": 0.767, | |
| "eval_wer": 0.6603700848111026, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 54.39, | |
| "learning_rate": 0.0001501730769230769, | |
| "loss": 0.8655, | |
| "step": 3100 | |
| }, | |
| { | |
| "epoch": 56.14, | |
| "learning_rate": 0.00014440384615384616, | |
| "loss": 0.8493, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 57.89, | |
| "learning_rate": 0.0001386923076923077, | |
| "loss": 0.839, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 59.65, | |
| "learning_rate": 0.00013292307692307692, | |
| "loss": 0.8204, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 61.4, | |
| "learning_rate": 0.00012715384615384615, | |
| "loss": 0.8223, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 61.4, | |
| "eval_loss": 0.8161018490791321, | |
| "eval_runtime": 34.0226, | |
| "eval_samples_per_second": 23.808, | |
| "eval_steps_per_second": 0.764, | |
| "eval_wer": 0.6968003084040093, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 63.16, | |
| "learning_rate": 0.00012138461538461537, | |
| "loss": 0.7998, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 64.91, | |
| "learning_rate": 0.00011561538461538461, | |
| "loss": 0.7853, | |
| "step": 3700 | |
| }, | |
| { | |
| "epoch": 66.67, | |
| "learning_rate": 0.00010984615384615384, | |
| "loss": 0.7698, | |
| "step": 3800 | |
| }, | |
| { | |
| "epoch": 68.42, | |
| "learning_rate": 0.00010407692307692307, | |
| "loss": 0.7631, | |
| "step": 3900 | |
| }, | |
| { | |
| "epoch": 70.18, | |
| "learning_rate": 9.83076923076923e-05, | |
| "loss": 0.7441, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 70.18, | |
| "eval_loss": 0.7056900858879089, | |
| "eval_runtime": 33.8841, | |
| "eval_samples_per_second": 23.905, | |
| "eval_steps_per_second": 0.767, | |
| "eval_wer": 0.6183500385505012, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 71.93, | |
| "learning_rate": 9.253846153846153e-05, | |
| "loss": 0.7343, | |
| "step": 4100 | |
| }, | |
| { | |
| "epoch": 73.68, | |
| "learning_rate": 8.676923076923076e-05, | |
| "loss": 0.7313, | |
| "step": 4200 | |
| }, | |
| { | |
| "epoch": 75.44, | |
| "learning_rate": 8.1e-05, | |
| "loss": 0.7176, | |
| "step": 4300 | |
| }, | |
| { | |
| "epoch": 77.19, | |
| "learning_rate": 7.523076923076922e-05, | |
| "loss": 0.7052, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 78.95, | |
| "learning_rate": 6.946153846153845e-05, | |
| "loss": 0.7011, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 78.95, | |
| "eval_loss": 0.7026883363723755, | |
| "eval_runtime": 33.7582, | |
| "eval_samples_per_second": 23.994, | |
| "eval_steps_per_second": 0.77, | |
| "eval_wer": 0.6023515805705474, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 80.7, | |
| "learning_rate": 6.369230769230768e-05, | |
| "loss": 0.6894, | |
| "step": 4600 | |
| }, | |
| { | |
| "epoch": 82.46, | |
| "learning_rate": 5.7923076923076925e-05, | |
| "loss": 0.6795, | |
| "step": 4700 | |
| }, | |
| { | |
| "epoch": 84.21, | |
| "learning_rate": 5.215384615384615e-05, | |
| "loss": 0.6771, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 85.96, | |
| "learning_rate": 4.6384615384615385e-05, | |
| "loss": 0.6624, | |
| "step": 4900 | |
| }, | |
| { | |
| "epoch": 87.72, | |
| "learning_rate": 4.061538461538461e-05, | |
| "loss": 0.6542, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 87.72, | |
| "eval_loss": 0.709155261516571, | |
| "eval_runtime": 34.8834, | |
| "eval_samples_per_second": 23.22, | |
| "eval_steps_per_second": 0.745, | |
| "eval_wer": 0.5979182729375482, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 89.47, | |
| "learning_rate": 3.484615384615384e-05, | |
| "loss": 0.6381, | |
| "step": 5100 | |
| }, | |
| { | |
| "epoch": 91.23, | |
| "learning_rate": 2.9076923076923072e-05, | |
| "loss": 0.6305, | |
| "step": 5200 | |
| }, | |
| { | |
| "epoch": 92.98, | |
| "learning_rate": 2.3307692307692303e-05, | |
| "loss": 0.6228, | |
| "step": 5300 | |
| }, | |
| { | |
| "epoch": 94.74, | |
| "learning_rate": 1.7538461538461536e-05, | |
| "loss": 0.6169, | |
| "step": 5400 | |
| }, | |
| { | |
| "epoch": 96.49, | |
| "learning_rate": 1.1769230769230768e-05, | |
| "loss": 0.6081, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 96.49, | |
| "eval_loss": 0.7916743755340576, | |
| "eval_runtime": 34.4838, | |
| "eval_samples_per_second": 23.489, | |
| "eval_steps_per_second": 0.754, | |
| "eval_wer": 0.6324209714726291, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 98.25, | |
| "learning_rate": 5.999999999999999e-06, | |
| "loss": 0.6121, | |
| "step": 5600 | |
| }, | |
| { | |
| "epoch": 100.0, | |
| "learning_rate": 2.3076923076923075e-07, | |
| "loss": 0.609, | |
| "step": 5700 | |
| }, | |
| { | |
| "epoch": 100.0, | |
| "step": 5700, | |
| "total_flos": 2.485998352803604e+19, | |
| "train_loss": 1.1609217145150168, | |
| "train_runtime": 9604.0577, | |
| "train_samples_per_second": 18.68, | |
| "train_steps_per_second": 0.593 | |
| } | |
| ], | |
| "max_steps": 5700, | |
| "num_train_epochs": 100, | |
| "total_flos": 2.485998352803604e+19, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |