| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 5.0, | |
| "eval_steps": 500, | |
| "global_step": 9805, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.25497195308516063, | |
| "grad_norm": 1.5315192937850952, | |
| "learning_rate": 0.00010193679918450561, | |
| "loss": 3.8945, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.5099439061703213, | |
| "grad_norm": 1.6545052528381348, | |
| "learning_rate": 0.00019956935630099728, | |
| "loss": 0.5228, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.7649158592554819, | |
| "grad_norm": 1.1093895435333252, | |
| "learning_rate": 0.0001882366273798731, | |
| "loss": 0.3702, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_cer": 0.13674156347728234, | |
| "eval_loss": 0.287818968296051, | |
| "eval_runtime": 242.9432, | |
| "eval_samples_per_second": 92.038, | |
| "eval_steps_per_second": 11.505, | |
| "eval_wer": 0.3335439517042918, | |
| "step": 1961 | |
| }, | |
| { | |
| "epoch": 1.0198878123406425, | |
| "grad_norm": 0.6819602847099304, | |
| "learning_rate": 0.0001769038984587489, | |
| "loss": 0.3114, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.2748597654258031, | |
| "grad_norm": 0.7735158801078796, | |
| "learning_rate": 0.00016557116953762465, | |
| "loss": 0.2571, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.5298317185109638, | |
| "grad_norm": 0.7301546931266785, | |
| "learning_rate": 0.00015423844061650046, | |
| "loss": 0.2411, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.7848036715961244, | |
| "grad_norm": 0.5918829441070557, | |
| "learning_rate": 0.00014290571169537624, | |
| "loss": 0.2333, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_cer": 0.12185579510710033, | |
| "eval_loss": 0.23235103487968445, | |
| "eval_runtime": 242.1687, | |
| "eval_samples_per_second": 92.332, | |
| "eval_steps_per_second": 11.542, | |
| "eval_wer": 0.26528299554267426, | |
| "step": 3922 | |
| }, | |
| { | |
| "epoch": 2.039775624681285, | |
| "grad_norm": 0.7791718244552612, | |
| "learning_rate": 0.00013157298277425205, | |
| "loss": 0.2157, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 2.2947475777664454, | |
| "grad_norm": 0.5816757082939148, | |
| "learning_rate": 0.00012024025385312785, | |
| "loss": 0.1817, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 2.5497195308516063, | |
| "grad_norm": 0.42094656825065613, | |
| "learning_rate": 0.00010890752493200362, | |
| "loss": 0.1803, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 2.804691483936767, | |
| "grad_norm": 0.515943706035614, | |
| "learning_rate": 9.757479601087942e-05, | |
| "loss": 0.172, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_cer": 0.11621454523249779, | |
| "eval_loss": 0.213576540350914, | |
| "eval_runtime": 242.2848, | |
| "eval_samples_per_second": 92.288, | |
| "eval_steps_per_second": 11.536, | |
| "eval_wer": 0.246385720074449, | |
| "step": 5883 | |
| }, | |
| { | |
| "epoch": 3.0596634370219276, | |
| "grad_norm": 0.397499680519104, | |
| "learning_rate": 8.624206708975522e-05, | |
| "loss": 0.1656, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 3.3146353901070884, | |
| "grad_norm": 0.5948652625083923, | |
| "learning_rate": 7.490933816863101e-05, | |
| "loss": 0.1388, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 3.569607343192249, | |
| "grad_norm": 0.5423580408096313, | |
| "learning_rate": 6.35766092475068e-05, | |
| "loss": 0.1383, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 3.8245792962774097, | |
| "grad_norm": 1.2003000974655151, | |
| "learning_rate": 5.22438803263826e-05, | |
| "loss": 0.1331, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_cer": 0.11265216708611207, | |
| "eval_loss": 0.20429827272891998, | |
| "eval_runtime": 242.3631, | |
| "eval_samples_per_second": 92.258, | |
| "eval_steps_per_second": 11.532, | |
| "eval_wer": 0.22865256897179317, | |
| "step": 7844 | |
| }, | |
| { | |
| "epoch": 4.07955124936257, | |
| "grad_norm": 0.8045985102653503, | |
| "learning_rate": 4.091115140525839e-05, | |
| "loss": 0.123, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 4.3345232024477305, | |
| "grad_norm": 0.402927964925766, | |
| "learning_rate": 2.957842248413418e-05, | |
| "loss": 0.1052, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 4.589495155532891, | |
| "grad_norm": 0.42731374502182007, | |
| "learning_rate": 1.8245693563009974e-05, | |
| "loss": 0.1054, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 4.844467108618052, | |
| "grad_norm": 0.2811175286769867, | |
| "learning_rate": 6.912964641885767e-06, | |
| "loss": 0.1018, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_cer": 0.10983672160892569, | |
| "eval_loss": 0.2057330459356308, | |
| "eval_runtime": 270.0068, | |
| "eval_samples_per_second": 82.813, | |
| "eval_steps_per_second": 10.352, | |
| "eval_wer": 0.21943229191706215, | |
| "step": 9805 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "step": 9805, | |
| "total_flos": 5.6911698946882765e+19, | |
| "train_loss": 0.39030011520891517, | |
| "train_runtime": 4855.0155, | |
| "train_samples_per_second": 64.613, | |
| "train_steps_per_second": 2.02 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 9805, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 5, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 5.6911698946882765e+19, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |