| { | |
| "best_metric": 18.938284039923083, | |
| "best_model_checkpoint": "./checkpoint-1000", | |
| "epoch": 9.7799511002445, | |
| "eval_steps": 1000, | |
| "global_step": 1000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.24, | |
| "grad_norm": 39.86528778076172, | |
| "learning_rate": 5.000000000000001e-07, | |
| "loss": 2.0555, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "grad_norm": Infinity, | |
| "learning_rate": 9.800000000000001e-07, | |
| "loss": 1.5219, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "grad_norm": 6.200109958648682, | |
| "learning_rate": 1.48e-06, | |
| "loss": 1.0167, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "grad_norm": 5.486103057861328, | |
| "learning_rate": 1.98e-06, | |
| "loss": 0.7299, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.22, | |
| "grad_norm": 5.532894134521484, | |
| "learning_rate": 2.4800000000000004e-06, | |
| "loss": 0.6318, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 1.47, | |
| "grad_norm": 4.895535945892334, | |
| "learning_rate": 2.9800000000000003e-06, | |
| "loss": 0.5503, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.71, | |
| "grad_norm": 5.065937519073486, | |
| "learning_rate": 3.48e-06, | |
| "loss": 0.4999, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 1.96, | |
| "grad_norm": 4.807600498199463, | |
| "learning_rate": 3.980000000000001e-06, | |
| "loss": 0.4457, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 2.2, | |
| "grad_norm": 4.568106174468994, | |
| "learning_rate": 4.48e-06, | |
| "loss": 0.3779, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 2.44, | |
| "grad_norm": 4.705833911895752, | |
| "learning_rate": 4.980000000000001e-06, | |
| "loss": 0.3332, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 2.69, | |
| "grad_norm": 4.315070629119873, | |
| "learning_rate": 5.480000000000001e-06, | |
| "loss": 0.2858, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 2.93, | |
| "grad_norm": 2.660728693008423, | |
| "learning_rate": 5.98e-06, | |
| "loss": 0.2228, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 3.18, | |
| "grad_norm": 2.5569632053375244, | |
| "learning_rate": 6.480000000000001e-06, | |
| "loss": 0.1766, | |
| "step": 325 | |
| }, | |
| { | |
| "epoch": 3.42, | |
| "grad_norm": 2.2663660049438477, | |
| "learning_rate": 6.98e-06, | |
| "loss": 0.1475, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 3.67, | |
| "grad_norm": 2.5257091522216797, | |
| "learning_rate": 7.48e-06, | |
| "loss": 0.1492, | |
| "step": 375 | |
| }, | |
| { | |
| "epoch": 3.91, | |
| "grad_norm": 2.3954405784606934, | |
| "learning_rate": 7.980000000000002e-06, | |
| "loss": 0.142, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 4.16, | |
| "grad_norm": 2.181328296661377, | |
| "learning_rate": 8.48e-06, | |
| "loss": 0.1125, | |
| "step": 425 | |
| }, | |
| { | |
| "epoch": 4.4, | |
| "grad_norm": 1.9877594709396362, | |
| "learning_rate": 8.98e-06, | |
| "loss": 0.0908, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 4.65, | |
| "grad_norm": 1.8853321075439453, | |
| "learning_rate": 9.48e-06, | |
| "loss": 0.0892, | |
| "step": 475 | |
| }, | |
| { | |
| "epoch": 4.89, | |
| "grad_norm": 2.3802549839019775, | |
| "learning_rate": 9.980000000000001e-06, | |
| "loss": 0.091, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 5.13, | |
| "grad_norm": 1.3576879501342773, | |
| "learning_rate": 9.946666666666667e-06, | |
| "loss": 0.0713, | |
| "step": 525 | |
| }, | |
| { | |
| "epoch": 5.38, | |
| "grad_norm": 2.4532103538513184, | |
| "learning_rate": 9.891111111111113e-06, | |
| "loss": 0.0534, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 5.62, | |
| "grad_norm": 1.3736106157302856, | |
| "learning_rate": 9.835555555555556e-06, | |
| "loss": 0.0512, | |
| "step": 575 | |
| }, | |
| { | |
| "epoch": 5.87, | |
| "grad_norm": 2.0095458030700684, | |
| "learning_rate": 9.780000000000001e-06, | |
| "loss": 0.0571, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 6.11, | |
| "grad_norm": 1.2924531698226929, | |
| "learning_rate": 9.724444444444445e-06, | |
| "loss": 0.0453, | |
| "step": 625 | |
| }, | |
| { | |
| "epoch": 6.36, | |
| "grad_norm": 1.0012321472167969, | |
| "learning_rate": 9.66888888888889e-06, | |
| "loss": 0.0292, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 6.6, | |
| "grad_norm": 1.4161434173583984, | |
| "learning_rate": 9.613333333333335e-06, | |
| "loss": 0.0325, | |
| "step": 675 | |
| }, | |
| { | |
| "epoch": 6.85, | |
| "grad_norm": 5.784367084503174, | |
| "learning_rate": 9.557777777777777e-06, | |
| "loss": 0.031, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 7.09, | |
| "grad_norm": 0.8382102251052856, | |
| "learning_rate": 9.502222222222223e-06, | |
| "loss": 0.0247, | |
| "step": 725 | |
| }, | |
| { | |
| "epoch": 7.33, | |
| "grad_norm": 1.2963491678237915, | |
| "learning_rate": 9.446666666666667e-06, | |
| "loss": 0.0162, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 7.58, | |
| "grad_norm": 1.7834402322769165, | |
| "learning_rate": 9.391111111111111e-06, | |
| "loss": 0.0175, | |
| "step": 775 | |
| }, | |
| { | |
| "epoch": 7.82, | |
| "grad_norm": 0.9083292484283447, | |
| "learning_rate": 9.335555555555557e-06, | |
| "loss": 0.0193, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 8.07, | |
| "grad_norm": 0.5552634596824646, | |
| "learning_rate": 9.280000000000001e-06, | |
| "loss": 0.0157, | |
| "step": 825 | |
| }, | |
| { | |
| "epoch": 8.31, | |
| "grad_norm": 1.1231069564819336, | |
| "learning_rate": 9.224444444444445e-06, | |
| "loss": 0.0105, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 8.56, | |
| "grad_norm": 1.206103801727295, | |
| "learning_rate": 9.168888888888889e-06, | |
| "loss": 0.0109, | |
| "step": 875 | |
| }, | |
| { | |
| "epoch": 8.8, | |
| "grad_norm": 0.8872191309928894, | |
| "learning_rate": 9.113333333333335e-06, | |
| "loss": 0.0126, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 9.05, | |
| "grad_norm": 0.7421383261680603, | |
| "learning_rate": 9.057777777777779e-06, | |
| "loss": 0.0107, | |
| "step": 925 | |
| }, | |
| { | |
| "epoch": 9.29, | |
| "grad_norm": 0.7581607103347778, | |
| "learning_rate": 9.002222222222223e-06, | |
| "loss": 0.006, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 9.54, | |
| "grad_norm": 0.6848894953727722, | |
| "learning_rate": 8.946666666666669e-06, | |
| "loss": 0.006, | |
| "step": 975 | |
| }, | |
| { | |
| "epoch": 9.78, | |
| "grad_norm": 1.044122576713562, | |
| "learning_rate": 8.891111111111111e-06, | |
| "loss": 0.0067, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 9.78, | |
| "eval_loss": 0.41375118494033813, | |
| "eval_runtime": 1461.809, | |
| "eval_samples_per_second": 1.98, | |
| "eval_steps_per_second": 0.495, | |
| "eval_wer": 18.938284039923083, | |
| "step": 1000 | |
| } | |
| ], | |
| "logging_steps": 25, | |
| "max_steps": 5000, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 50, | |
| "save_steps": 1000, | |
| "total_flos": 1.845907654606848e+19, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |