| { | |
| "best_metric": 0.0, | |
| "best_model_checkpoint": "D:\\development\\whisper-finetune\\output\\checkpoint-1000", | |
| "epoch": 15000.0, | |
| "eval_steps": 1000, | |
| "global_step": 15000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 500.0, | |
| "grad_norm": 0.000682959274854511, | |
| "learning_rate": 9.920000000000002e-06, | |
| "loss": 0.5312, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1000.0, | |
| "grad_norm": 0.00020609197963494807, | |
| "learning_rate": 9.65793103448276e-06, | |
| "loss": 0.0, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1000.0, | |
| "eval_loss": 2.3341947326116497e-06, | |
| "eval_runtime": 0.6984, | |
| "eval_samples_per_second": 4.295, | |
| "eval_steps_per_second": 1.432, | |
| "eval_wer": 0.0, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1500.0, | |
| "grad_norm": 0.0001350301317870617, | |
| "learning_rate": 9.313103448275864e-06, | |
| "loss": 0.0, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 2000.0, | |
| "grad_norm": 0.00011808969429694116, | |
| "learning_rate": 8.968275862068966e-06, | |
| "loss": 0.0, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2000.0, | |
| "eval_loss": 1.0998018069585669e-06, | |
| "eval_runtime": 0.7217, | |
| "eval_samples_per_second": 4.157, | |
| "eval_steps_per_second": 1.386, | |
| "eval_wer": 0.0, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2500.0, | |
| "grad_norm": 8.354395686183125e-05, | |
| "learning_rate": 8.623448275862069e-06, | |
| "loss": 0.0, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 3000.0, | |
| "grad_norm": 8.122230065055192e-05, | |
| "learning_rate": 8.278620689655173e-06, | |
| "loss": 0.0, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 3000.0, | |
| "eval_loss": 6.614192784581974e-07, | |
| "eval_runtime": 0.6797, | |
| "eval_samples_per_second": 4.414, | |
| "eval_steps_per_second": 1.471, | |
| "eval_wer": 0.0, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 3500.0, | |
| "grad_norm": 8.345582318725064e-05, | |
| "learning_rate": 7.933793103448275e-06, | |
| "loss": 0.0, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 4000.0, | |
| "grad_norm": 8.580592839280143e-05, | |
| "learning_rate": 7.58896551724138e-06, | |
| "loss": 0.0, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 4000.0, | |
| "eval_loss": 6.152737341835746e-07, | |
| "eval_runtime": 0.6701, | |
| "eval_samples_per_second": 4.477, | |
| "eval_steps_per_second": 1.492, | |
| "eval_wer": 0.0, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 4500.0, | |
| "grad_norm": 8.349263953277841e-05, | |
| "learning_rate": 7.244137931034483e-06, | |
| "loss": 0.0, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 5000.0, | |
| "grad_norm": 8.662583422847092e-05, | |
| "learning_rate": 6.899310344827586e-06, | |
| "loss": 0.0, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 5000.0, | |
| "eval_loss": 5.268281597636815e-07, | |
| "eval_runtime": 0.6852, | |
| "eval_samples_per_second": 4.378, | |
| "eval_steps_per_second": 1.459, | |
| "eval_wer": 0.0, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 5500.0, | |
| "grad_norm": 9.030300134327263e-05, | |
| "learning_rate": 6.5544827586206905e-06, | |
| "loss": 0.0, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 6000.0, | |
| "grad_norm": 9.486892668064684e-05, | |
| "learning_rate": 6.209655172413793e-06, | |
| "loss": 0.0, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 6000.0, | |
| "eval_loss": 5.345190743355488e-07, | |
| "eval_runtime": 0.6873, | |
| "eval_samples_per_second": 4.365, | |
| "eval_steps_per_second": 1.455, | |
| "eval_wer": 0.0, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 6500.0, | |
| "grad_norm": 6.878418935230002e-05, | |
| "learning_rate": 5.864827586206897e-06, | |
| "loss": 0.0, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 7000.0, | |
| "grad_norm": 0.0008927604649215937, | |
| "learning_rate": 5.5200000000000005e-06, | |
| "loss": 0.0, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 7000.0, | |
| "eval_loss": 2.384185791015625e-07, | |
| "eval_runtime": 0.6998, | |
| "eval_samples_per_second": 4.287, | |
| "eval_steps_per_second": 1.429, | |
| "eval_wer": 0.0, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 7500.0, | |
| "grad_norm": 0.0001970465964404866, | |
| "learning_rate": 5.175172413793104e-06, | |
| "loss": 0.0, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 8000.0, | |
| "grad_norm": 0.00019531552970875055, | |
| "learning_rate": 4.830344827586207e-06, | |
| "loss": 0.0, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 8000.0, | |
| "eval_loss": 1.5766390504268202e-07, | |
| "eval_runtime": 0.7272, | |
| "eval_samples_per_second": 4.126, | |
| "eval_steps_per_second": 1.375, | |
| "eval_wer": 0.0, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 8500.0, | |
| "grad_norm": 8.719138713786379e-05, | |
| "learning_rate": 4.4855172413793105e-06, | |
| "loss": 0.0, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 9000.0, | |
| "grad_norm": 9.791918273549527e-05, | |
| "learning_rate": 4.140689655172414e-06, | |
| "loss": 0.0, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 9000.0, | |
| "eval_loss": 3.076368670917873e-07, | |
| "eval_runtime": 0.7114, | |
| "eval_samples_per_second": 4.217, | |
| "eval_steps_per_second": 1.406, | |
| "eval_wer": 0.0, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 9500.0, | |
| "grad_norm": 0.0005698113818652928, | |
| "learning_rate": 3.795862068965517e-06, | |
| "loss": 0.0, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 10000.0, | |
| "grad_norm": 0.00010731745715020224, | |
| "learning_rate": 3.4510344827586214e-06, | |
| "loss": 0.0, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 10000.0, | |
| "eval_loss": 5.7681912579710115e-08, | |
| "eval_runtime": 0.7029, | |
| "eval_samples_per_second": 4.268, | |
| "eval_steps_per_second": 1.423, | |
| "eval_wer": 0.0, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 10500.0, | |
| "grad_norm": 0.00012043450260534883, | |
| "learning_rate": 3.1062068965517243e-06, | |
| "loss": 0.0, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 11000.0, | |
| "grad_norm": 0.00018969921802636236, | |
| "learning_rate": 2.7613793103448276e-06, | |
| "loss": 0.0, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 11000.0, | |
| "eval_loss": 1.3074567561943695e-07, | |
| "eval_runtime": 0.698, | |
| "eval_samples_per_second": 4.298, | |
| "eval_steps_per_second": 1.433, | |
| "eval_wer": 0.0, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 11500.0, | |
| "grad_norm": 0.00011485564755275846, | |
| "learning_rate": 2.4165517241379314e-06, | |
| "loss": 0.0, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 12000.0, | |
| "grad_norm": 9.144964133156464e-05, | |
| "learning_rate": 2.0717241379310347e-06, | |
| "loss": 0.0, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 12000.0, | |
| "eval_loss": 2.1150034967831743e-07, | |
| "eval_runtime": 0.6885, | |
| "eval_samples_per_second": 4.357, | |
| "eval_steps_per_second": 1.452, | |
| "eval_wer": 0.0, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 12500.0, | |
| "grad_norm": 8.384387183468789e-05, | |
| "learning_rate": 1.7268965517241382e-06, | |
| "loss": 0.0, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 13000.0, | |
| "grad_norm": 8.791466825641692e-05, | |
| "learning_rate": 1.3820689655172416e-06, | |
| "loss": 0.0, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 13000.0, | |
| "eval_loss": 9.998198180483087e-08, | |
| "eval_runtime": 0.6646, | |
| "eval_samples_per_second": 4.514, | |
| "eval_steps_per_second": 1.505, | |
| "eval_wer": 0.0, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 13500.0, | |
| "grad_norm": 6.821281567681581e-05, | |
| "learning_rate": 1.0372413793103449e-06, | |
| "loss": 0.0, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 14000.0, | |
| "grad_norm": 5.373560270527378e-05, | |
| "learning_rate": 6.924137931034483e-07, | |
| "loss": 0.0, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 14000.0, | |
| "eval_loss": 1.3459113290537061e-07, | |
| "eval_runtime": 0.6976, | |
| "eval_samples_per_second": 4.301, | |
| "eval_steps_per_second": 1.434, | |
| "eval_wer": 0.0, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 14500.0, | |
| "grad_norm": 8.191769302356988e-05, | |
| "learning_rate": 3.475862068965517e-07, | |
| "loss": 0.0, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 15000.0, | |
| "grad_norm": 5.6634093198226765e-05, | |
| "learning_rate": 2.758620689655173e-09, | |
| "loss": 0.0, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 15000.0, | |
| "eval_loss": 1.3843659019130428e-07, | |
| "eval_runtime": 0.6813, | |
| "eval_samples_per_second": 4.403, | |
| "eval_steps_per_second": 1.468, | |
| "eval_wer": 0.0, | |
| "step": 15000 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 15000, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 15000, | |
| "save_steps": 1000, | |
| "total_flos": 2.9186998272e+18, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |