| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 10.0, | |
| "eval_steps": 500, | |
| "global_step": 4040, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.12, | |
| "learning_rate": 5.000000000000001e-07, | |
| "loss": 10.9669, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "learning_rate": 1.0000000000000002e-06, | |
| "loss": 10.7574, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "learning_rate": 1.5e-06, | |
| "loss": 10.4847, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "learning_rate": 2.0000000000000003e-06, | |
| "loss": 9.3795, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "learning_rate": 2.5e-06, | |
| "loss": 7.4952, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "learning_rate": 3e-06, | |
| "loss": 6.3533, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.87, | |
| "learning_rate": 3.5e-06, | |
| "loss": 5.356, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "learning_rate": 4.000000000000001e-06, | |
| "loss": 5.0318, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_cer": 0.9888512643814729, | |
| "eval_loss": 4.29986047744751, | |
| "eval_runtime": 38.3788, | |
| "eval_samples_per_second": 168.426, | |
| "eval_steps_per_second": 10.527, | |
| "eval_wer": 0.9797794777861581, | |
| "step": 404 | |
| }, | |
| { | |
| "epoch": 1.11, | |
| "learning_rate": 4.5e-06, | |
| "loss": 4.7526, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 1.24, | |
| "learning_rate": 5e-06, | |
| "loss": 4.5027, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.36, | |
| "learning_rate": 5.500000000000001e-06, | |
| "loss": 4.3223, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 1.49, | |
| "learning_rate": 6e-06, | |
| "loss": 4.0849, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.61, | |
| "learning_rate": 6.5000000000000004e-06, | |
| "loss": 3.9545, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 1.73, | |
| "learning_rate": 7e-06, | |
| "loss": 3.7431, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 1.86, | |
| "learning_rate": 7.500000000000001e-06, | |
| "loss": 3.6413, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 1.98, | |
| "learning_rate": 8.000000000000001e-06, | |
| "loss": 3.5113, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_cer": 0.9888512643814729, | |
| "eval_loss": 3.3288848400115967, | |
| "eval_runtime": 39.0341, | |
| "eval_samples_per_second": 165.599, | |
| "eval_steps_per_second": 10.35, | |
| "eval_wer": 0.9797794777861581, | |
| "step": 808 | |
| }, | |
| { | |
| "epoch": 2.1, | |
| "learning_rate": 8.5e-06, | |
| "loss": 3.3907, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 2.23, | |
| "learning_rate": 9e-06, | |
| "loss": 3.2836, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.35, | |
| "learning_rate": 9.5e-06, | |
| "loss": 3.1635, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 2.48, | |
| "learning_rate": 1e-05, | |
| "loss": 3.0437, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.6, | |
| "learning_rate": 9.835526315789474e-06, | |
| "loss": 2.9567, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 2.72, | |
| "learning_rate": 9.671052631578948e-06, | |
| "loss": 2.8632, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 2.85, | |
| "learning_rate": 9.506578947368423e-06, | |
| "loss": 2.8078, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 2.97, | |
| "learning_rate": 9.342105263157895e-06, | |
| "loss": 2.7536, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_cer": 0.9888512643814729, | |
| "eval_loss": 2.700655460357666, | |
| "eval_runtime": 38.3318, | |
| "eval_samples_per_second": 168.633, | |
| "eval_steps_per_second": 10.54, | |
| "eval_wer": 0.9797794777861581, | |
| "step": 1212 | |
| }, | |
| { | |
| "epoch": 3.09, | |
| "learning_rate": 9.17763157894737e-06, | |
| "loss": 2.7127, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 3.22, | |
| "learning_rate": 9.013157894736843e-06, | |
| "loss": 2.6749, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 3.34, | |
| "learning_rate": 8.848684210526316e-06, | |
| "loss": 2.6572, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 3.47, | |
| "learning_rate": 8.68421052631579e-06, | |
| "loss": 2.6236, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 3.59, | |
| "learning_rate": 8.519736842105265e-06, | |
| "loss": 2.6007, | |
| "step": 1450 | |
| }, | |
| { | |
| "epoch": 3.71, | |
| "learning_rate": 8.355263157894737e-06, | |
| "loss": 2.5653, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 3.84, | |
| "learning_rate": 8.19078947368421e-06, | |
| "loss": 2.5333, | |
| "step": 1550 | |
| }, | |
| { | |
| "epoch": 3.96, | |
| "learning_rate": 8.026315789473685e-06, | |
| "loss": 2.4826, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_cer": 0.9888512643814729, | |
| "eval_loss": 2.373181104660034, | |
| "eval_runtime": 37.8617, | |
| "eval_samples_per_second": 170.727, | |
| "eval_steps_per_second": 10.67, | |
| "eval_wer": 0.9797794777861581, | |
| "step": 1616 | |
| }, | |
| { | |
| "epoch": 4.08, | |
| "learning_rate": 7.86184210526316e-06, | |
| "loss": 2.4176, | |
| "step": 1650 | |
| }, | |
| { | |
| "epoch": 4.21, | |
| "learning_rate": 7.697368421052632e-06, | |
| "loss": 2.371, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 4.33, | |
| "learning_rate": 7.532894736842106e-06, | |
| "loss": 2.302, | |
| "step": 1750 | |
| }, | |
| { | |
| "epoch": 4.46, | |
| "learning_rate": 7.368421052631579e-06, | |
| "loss": 2.2307, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 4.58, | |
| "learning_rate": 7.203947368421054e-06, | |
| "loss": 2.1823, | |
| "step": 1850 | |
| }, | |
| { | |
| "epoch": 4.7, | |
| "learning_rate": 7.0394736842105274e-06, | |
| "loss": 2.157, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 4.83, | |
| "learning_rate": 6.875e-06, | |
| "loss": 2.1087, | |
| "step": 1950 | |
| }, | |
| { | |
| "epoch": 4.95, | |
| "learning_rate": 6.710526315789474e-06, | |
| "loss": 2.0642, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_cer": 0.9888314620091487, | |
| "eval_loss": 1.9164613485336304, | |
| "eval_runtime": 38.0086, | |
| "eval_samples_per_second": 170.067, | |
| "eval_steps_per_second": 10.629, | |
| "eval_wer": 0.9797794777861581, | |
| "step": 2020 | |
| }, | |
| { | |
| "epoch": 5.07, | |
| "learning_rate": 6.5460526315789476e-06, | |
| "loss": 2.0318, | |
| "step": 2050 | |
| }, | |
| { | |
| "epoch": 5.2, | |
| "learning_rate": 6.381578947368422e-06, | |
| "loss": 2.0039, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 5.32, | |
| "learning_rate": 6.217105263157896e-06, | |
| "loss": 1.97, | |
| "step": 2150 | |
| }, | |
| { | |
| "epoch": 5.45, | |
| "learning_rate": 6.0526315789473685e-06, | |
| "loss": 1.9502, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 5.57, | |
| "learning_rate": 5.888157894736842e-06, | |
| "loss": 1.9235, | |
| "step": 2250 | |
| }, | |
| { | |
| "epoch": 5.69, | |
| "learning_rate": 5.723684210526316e-06, | |
| "loss": 1.8816, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 5.82, | |
| "learning_rate": 5.55921052631579e-06, | |
| "loss": 1.8675, | |
| "step": 2350 | |
| }, | |
| { | |
| "epoch": 5.94, | |
| "learning_rate": 5.394736842105264e-06, | |
| "loss": 1.834, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_cer": 0.9463949781183786, | |
| "eval_loss": 1.6738649606704712, | |
| "eval_runtime": 38.6983, | |
| "eval_samples_per_second": 167.036, | |
| "eval_steps_per_second": 10.44, | |
| "eval_wer": 0.9504004597205761, | |
| "step": 2424 | |
| }, | |
| { | |
| "epoch": 6.06, | |
| "learning_rate": 5.230263157894737e-06, | |
| "loss": 1.8211, | |
| "step": 2450 | |
| }, | |
| { | |
| "epoch": 6.19, | |
| "learning_rate": 5.0657894736842104e-06, | |
| "loss": 1.7972, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 6.31, | |
| "learning_rate": 4.901315789473685e-06, | |
| "loss": 1.7812, | |
| "step": 2550 | |
| }, | |
| { | |
| "epoch": 6.44, | |
| "learning_rate": 4.736842105263158e-06, | |
| "loss": 1.7627, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 6.56, | |
| "learning_rate": 4.572368421052632e-06, | |
| "loss": 1.7391, | |
| "step": 2650 | |
| }, | |
| { | |
| "epoch": 6.68, | |
| "learning_rate": 4.407894736842105e-06, | |
| "loss": 1.7241, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 6.81, | |
| "learning_rate": 4.2434210526315796e-06, | |
| "loss": 1.7089, | |
| "step": 2750 | |
| }, | |
| { | |
| "epoch": 6.93, | |
| "learning_rate": 4.078947368421053e-06, | |
| "loss": 1.6869, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_cer": 0.7864710192281035, | |
| "eval_loss": 1.4651445150375366, | |
| "eval_runtime": 38.8273, | |
| "eval_samples_per_second": 166.481, | |
| "eval_steps_per_second": 10.405, | |
| "eval_wer": 0.8239413856265488, | |
| "step": 2828 | |
| }, | |
| { | |
| "epoch": 7.05, | |
| "learning_rate": 3.914473684210527e-06, | |
| "loss": 1.6732, | |
| "step": 2850 | |
| }, | |
| { | |
| "epoch": 7.18, | |
| "learning_rate": 3.7500000000000005e-06, | |
| "loss": 1.6575, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 7.3, | |
| "learning_rate": 3.5855263157894737e-06, | |
| "loss": 1.6262, | |
| "step": 2950 | |
| }, | |
| { | |
| "epoch": 7.43, | |
| "learning_rate": 3.421052631578948e-06, | |
| "loss": 1.635, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 7.55, | |
| "learning_rate": 3.256578947368421e-06, | |
| "loss": 1.6008, | |
| "step": 3050 | |
| }, | |
| { | |
| "epoch": 7.67, | |
| "learning_rate": 3.092105263157895e-06, | |
| "loss": 1.5957, | |
| "step": 3100 | |
| }, | |
| { | |
| "epoch": 7.8, | |
| "learning_rate": 2.927631578947369e-06, | |
| "loss": 1.5809, | |
| "step": 3150 | |
| }, | |
| { | |
| "epoch": 7.92, | |
| "learning_rate": 2.7631578947368424e-06, | |
| "loss": 1.5734, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_cer": 0.693914730984772, | |
| "eval_loss": 1.3266735076904297, | |
| "eval_runtime": 37.9957, | |
| "eval_samples_per_second": 170.125, | |
| "eval_steps_per_second": 10.633, | |
| "eval_wer": 0.7439571885213518, | |
| "step": 3232 | |
| }, | |
| { | |
| "epoch": 8.04, | |
| "learning_rate": 2.598684210526316e-06, | |
| "loss": 1.5604, | |
| "step": 3250 | |
| }, | |
| { | |
| "epoch": 8.17, | |
| "learning_rate": 2.4342105263157898e-06, | |
| "loss": 1.5457, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 8.29, | |
| "learning_rate": 2.2697368421052634e-06, | |
| "loss": 1.5277, | |
| "step": 3350 | |
| }, | |
| { | |
| "epoch": 8.42, | |
| "learning_rate": 2.105263157894737e-06, | |
| "loss": 1.53, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 8.54, | |
| "learning_rate": 1.9407894736842107e-06, | |
| "loss": 1.5365, | |
| "step": 3450 | |
| }, | |
| { | |
| "epoch": 8.66, | |
| "learning_rate": 1.7763157894736844e-06, | |
| "loss": 1.5237, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 8.79, | |
| "learning_rate": 1.611842105263158e-06, | |
| "loss": 1.5106, | |
| "step": 3550 | |
| }, | |
| { | |
| "epoch": 8.91, | |
| "learning_rate": 1.4473684210526317e-06, | |
| "loss": 1.5052, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_cer": 0.6230816451810927, | |
| "eval_loss": 1.233142614364624, | |
| "eval_runtime": 39.1396, | |
| "eval_samples_per_second": 165.152, | |
| "eval_steps_per_second": 10.322, | |
| "eval_wer": 0.7045217828538591, | |
| "step": 3636 | |
| }, | |
| { | |
| "epoch": 9.03, | |
| "learning_rate": 1.2828947368421055e-06, | |
| "loss": 1.4932, | |
| "step": 3650 | |
| }, | |
| { | |
| "epoch": 9.16, | |
| "learning_rate": 1.118421052631579e-06, | |
| "loss": 1.4764, | |
| "step": 3700 | |
| }, | |
| { | |
| "epoch": 9.28, | |
| "learning_rate": 9.539473684210528e-07, | |
| "loss": 1.4474, | |
| "step": 3750 | |
| }, | |
| { | |
| "epoch": 9.41, | |
| "learning_rate": 7.894736842105263e-07, | |
| "loss": 1.4755, | |
| "step": 3800 | |
| }, | |
| { | |
| "epoch": 9.53, | |
| "learning_rate": 6.25e-07, | |
| "loss": 1.481, | |
| "step": 3850 | |
| }, | |
| { | |
| "epoch": 9.65, | |
| "learning_rate": 4.605263157894737e-07, | |
| "loss": 1.4635, | |
| "step": 3900 | |
| }, | |
| { | |
| "epoch": 9.78, | |
| "learning_rate": 2.9605263157894736e-07, | |
| "loss": 1.4576, | |
| "step": 3950 | |
| }, | |
| { | |
| "epoch": 9.9, | |
| "learning_rate": 1.3157894736842107e-07, | |
| "loss": 1.4573, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_cer": 0.5793976118338977, | |
| "eval_loss": 1.198919653892517, | |
| "eval_runtime": 38.6365, | |
| "eval_samples_per_second": 167.303, | |
| "eval_steps_per_second": 10.456, | |
| "eval_wer": 0.680063211579212, | |
| "step": 4040 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 4040, | |
| "total_flos": 2.1546455399489285e+18, | |
| "train_loss": 2.8324050827781755, | |
| "train_runtime": 8680.4972, | |
| "train_samples_per_second": 29.786, | |
| "train_steps_per_second": 0.465 | |
| } | |
| ], | |
| "logging_steps": 50, | |
| "max_steps": 4040, | |
| "num_train_epochs": 10, | |
| "save_steps": 500, | |
| "total_flos": 2.1546455399489285e+18, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |