| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 10.0, | |
| "global_step": 3480, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.29, | |
| "learning_rate": 3.7499999999999997e-06, | |
| "loss": 47.1781, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "learning_rate": 7.499999999999999e-06, | |
| "loss": 33.2186, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "learning_rate": 1.1249999999999999e-05, | |
| "loss": 26.4407, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 1.15, | |
| "learning_rate": 1.4999999999999999e-05, | |
| "loss": 23.1487, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.44, | |
| "learning_rate": 1.875e-05, | |
| "loss": 19.558, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.44, | |
| "eval_cer": 1.0, | |
| "eval_loss": 19.4094295501709, | |
| "eval_runtime": 25.117, | |
| "eval_samples_per_second": 18.155, | |
| "eval_steps_per_second": 2.269, | |
| "eval_wer": 1.0, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.72, | |
| "learning_rate": 2.2499999999999998e-05, | |
| "loss": 15.5015, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 2.01, | |
| "learning_rate": 2.6249999999999998e-05, | |
| "loss": 11.0351, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 2.3, | |
| "learning_rate": 2.9999999999999997e-05, | |
| "loss": 7.2384, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.59, | |
| "learning_rate": 3.375e-05, | |
| "loss": 5.2354, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.87, | |
| "learning_rate": 3.75e-05, | |
| "loss": 4.7968, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.87, | |
| "eval_cer": 1.0, | |
| "eval_loss": 4.782816410064697, | |
| "eval_runtime": 25.1318, | |
| "eval_samples_per_second": 18.144, | |
| "eval_steps_per_second": 2.268, | |
| "eval_wer": 1.0, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 3.16, | |
| "learning_rate": 4.125e-05, | |
| "loss": 4.7437, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 3.45, | |
| "learning_rate": 4.4999999999999996e-05, | |
| "loss": 4.654, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 3.74, | |
| "learning_rate": 4.875e-05, | |
| "loss": 4.6017, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 4.02, | |
| "learning_rate": 5.2499999999999995e-05, | |
| "loss": 4.5759, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 4.31, | |
| "learning_rate": 5.625e-05, | |
| "loss": 4.5125, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 4.31, | |
| "eval_cer": 0.9540229885057471, | |
| "eval_loss": 4.495882987976074, | |
| "eval_runtime": 25.1686, | |
| "eval_samples_per_second": 18.118, | |
| "eval_steps_per_second": 2.265, | |
| "eval_wer": 0.9990939293264874, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 4.6, | |
| "learning_rate": 5.9999999999999995e-05, | |
| "loss": 4.3867, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 4.89, | |
| "learning_rate": 6.374999999999999e-05, | |
| "loss": 4.304, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 5.17, | |
| "learning_rate": 6.75e-05, | |
| "loss": 4.261, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 5.46, | |
| "learning_rate": 7.125e-05, | |
| "loss": 4.2279, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 5.75, | |
| "learning_rate": 7.5e-05, | |
| "loss": 4.2202, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 5.75, | |
| "eval_cer": 0.8520311761927255, | |
| "eval_loss": 4.290519714355469, | |
| "eval_runtime": 24.9956, | |
| "eval_samples_per_second": 18.243, | |
| "eval_steps_per_second": 2.28, | |
| "eval_wer": 0.9922983992751434, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 6.03, | |
| "learning_rate": 6.993243243243243e-05, | |
| "loss": 4.184, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 6.32, | |
| "learning_rate": 6.486486486486486e-05, | |
| "loss": 4.1509, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 6.61, | |
| "learning_rate": 5.979729729729729e-05, | |
| "loss": 4.0915, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 6.9, | |
| "learning_rate": 5.4729729729729724e-05, | |
| "loss": 3.9577, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 7.18, | |
| "learning_rate": 4.966216216216216e-05, | |
| "loss": 3.7774, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 7.18, | |
| "eval_cer": 0.6651708392379153, | |
| "eval_loss": 3.2845523357391357, | |
| "eval_runtime": 25.0256, | |
| "eval_samples_per_second": 18.221, | |
| "eval_steps_per_second": 2.278, | |
| "eval_wer": 1.0356387798248263, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 7.47, | |
| "learning_rate": 4.4594594594594596e-05, | |
| "loss": 3.601, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 7.76, | |
| "learning_rate": 3.952702702702702e-05, | |
| "loss": 3.4439, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 8.05, | |
| "learning_rate": 3.4459459459459455e-05, | |
| "loss": 3.3193, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 8.33, | |
| "learning_rate": 2.9391891891891888e-05, | |
| "loss": 3.2091, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 8.62, | |
| "learning_rate": 2.4324324324324324e-05, | |
| "loss": 3.1418, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 8.62, | |
| "eval_cer": 0.5429459927570461, | |
| "eval_loss": 2.362445592880249, | |
| "eval_runtime": 25.1664, | |
| "eval_samples_per_second": 18.119, | |
| "eval_steps_per_second": 2.265, | |
| "eval_wer": 0.988221081244337, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 8.91, | |
| "learning_rate": 1.9256756756756753e-05, | |
| "loss": 3.0785, | |
| "step": 3100 | |
| }, | |
| { | |
| "epoch": 9.2, | |
| "learning_rate": 1.4189189189189189e-05, | |
| "loss": 3.0369, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 9.48, | |
| "learning_rate": 9.121621621621622e-06, | |
| "loss": 3.003, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 9.77, | |
| "learning_rate": 4.054054054054054e-06, | |
| "loss": 2.977, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 3480, | |
| "total_flos": 5.681228030880589e+19, | |
| "train_loss": 8.309904957913805, | |
| "train_runtime": 17807.6517, | |
| "train_samples_per_second": 12.501, | |
| "train_steps_per_second": 0.195 | |
| } | |
| ], | |
| "max_steps": 3480, | |
| "num_train_epochs": 10, | |
| "total_flos": 5.681228030880589e+19, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |