| { | |
| "best_metric": 0.9502835538752363, | |
| "best_model_checkpoint": "BengaliRegionalASR_finetune1/checkpoint-400", | |
| "epoch": 88.88888888888889, | |
| "eval_steps": 50, | |
| "global_step": 400, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 2.2222222222222223, | |
| "grad_norm": 8.39985179901123, | |
| "learning_rate": 2.0000000000000003e-06, | |
| "loss": 0.8375, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 4.444444444444445, | |
| "grad_norm": 4.356942176818848, | |
| "learning_rate": 4.000000000000001e-06, | |
| "loss": 0.7948, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 6.666666666666667, | |
| "grad_norm": 3.8019840717315674, | |
| "learning_rate": 6e-06, | |
| "loss": 0.7282, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 8.88888888888889, | |
| "grad_norm": 11.102900505065918, | |
| "learning_rate": 7.800000000000002e-06, | |
| "loss": 0.6569, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 11.11111111111111, | |
| "grad_norm": 12.99394416809082, | |
| "learning_rate": 9.600000000000001e-06, | |
| "loss": 0.6086, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 11.11111111111111, | |
| "eval_loss": 0.6732638478279114, | |
| "eval_runtime": 20.2721, | |
| "eval_samples_per_second": 4.193, | |
| "eval_steps_per_second": 0.099, | |
| "eval_wer": 1.1232850609756098, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 13.333333333333334, | |
| "grad_norm": 14.49991226196289, | |
| "learning_rate": 9.771428571428571e-06, | |
| "loss": 0.568, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 15.555555555555555, | |
| "grad_norm": 9.160536766052246, | |
| "learning_rate": 9.485714285714287e-06, | |
| "loss": 0.5212, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 17.77777777777778, | |
| "grad_norm": 5.1057634353637695, | |
| "learning_rate": 9.200000000000002e-06, | |
| "loss": 0.4796, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "grad_norm": 8.087906837463379, | |
| "learning_rate": 8.914285714285716e-06, | |
| "loss": 0.4458, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 22.22222222222222, | |
| "grad_norm": 6.61909294128418, | |
| "learning_rate": 8.628571428571429e-06, | |
| "loss": 0.4243, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 22.22222222222222, | |
| "eval_loss": 0.5663390159606934, | |
| "eval_runtime": 19.9974, | |
| "eval_samples_per_second": 4.251, | |
| "eval_steps_per_second": 0.1, | |
| "eval_wer": 1.068579027355623, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 24.444444444444443, | |
| "grad_norm": 4.257968902587891, | |
| "learning_rate": 8.342857142857143e-06, | |
| "loss": 0.4015, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 26.666666666666668, | |
| "grad_norm": 4.2625226974487305, | |
| "learning_rate": 8.057142857142857e-06, | |
| "loss": 0.3832, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 28.88888888888889, | |
| "grad_norm": 4.276031970977783, | |
| "learning_rate": 7.771428571428572e-06, | |
| "loss": 0.3557, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 31.11111111111111, | |
| "grad_norm": 12.266279220581055, | |
| "learning_rate": 7.485714285714286e-06, | |
| "loss": 0.3353, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 33.333333333333336, | |
| "grad_norm": 9.257742881774902, | |
| "learning_rate": 7.2000000000000005e-06, | |
| "loss": 0.3123, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 33.333333333333336, | |
| "eval_loss": 0.5672491788864136, | |
| "eval_runtime": 20.5784, | |
| "eval_samples_per_second": 4.131, | |
| "eval_steps_per_second": 0.097, | |
| "eval_wer": 0.9937488160636484, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 35.55555555555556, | |
| "grad_norm": 4.698246479034424, | |
| "learning_rate": 6.914285714285715e-06, | |
| "loss": 0.2906, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 37.77777777777778, | |
| "grad_norm": 4.475348472595215, | |
| "learning_rate": 6.628571428571428e-06, | |
| "loss": 0.2597, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 40.0, | |
| "grad_norm": 2.7412610054016113, | |
| "learning_rate": 6.342857142857143e-06, | |
| "loss": 0.2397, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 42.22222222222222, | |
| "grad_norm": 2.8001537322998047, | |
| "learning_rate": 6.057142857142858e-06, | |
| "loss": 0.2183, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 44.44444444444444, | |
| "grad_norm": 2.7077558040618896, | |
| "learning_rate": 5.771428571428572e-06, | |
| "loss": 0.2013, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 44.44444444444444, | |
| "eval_loss": 0.5929003953933716, | |
| "eval_runtime": 20.3009, | |
| "eval_samples_per_second": 4.187, | |
| "eval_steps_per_second": 0.099, | |
| "eval_wer": 0.9850046860356139, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 46.666666666666664, | |
| "grad_norm": 2.755575656890869, | |
| "learning_rate": 5.485714285714287e-06, | |
| "loss": 0.1848, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 48.888888888888886, | |
| "grad_norm": 3.3562865257263184, | |
| "learning_rate": 5.2e-06, | |
| "loss": 0.1697, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 51.111111111111114, | |
| "grad_norm": 6.205230236053467, | |
| "learning_rate": 4.9142857142857145e-06, | |
| "loss": 0.1577, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 53.333333333333336, | |
| "grad_norm": 4.571589946746826, | |
| "learning_rate": 4.628571428571429e-06, | |
| "loss": 0.148, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 55.55555555555556, | |
| "grad_norm": 4.679986476898193, | |
| "learning_rate": 4.342857142857143e-06, | |
| "loss": 0.1369, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 55.55555555555556, | |
| "eval_loss": 0.6405652761459351, | |
| "eval_runtime": 20.0641, | |
| "eval_samples_per_second": 4.236, | |
| "eval_steps_per_second": 0.1, | |
| "eval_wer": 0.9656160458452722, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 57.77777777777778, | |
| "grad_norm": 3.558831214904785, | |
| "learning_rate": 4.057142857142858e-06, | |
| "loss": 0.1251, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 60.0, | |
| "grad_norm": 2.806180000305176, | |
| "learning_rate": 3.771428571428572e-06, | |
| "loss": 0.1162, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 62.22222222222222, | |
| "grad_norm": 3.5792713165283203, | |
| "learning_rate": 3.4857142857142863e-06, | |
| "loss": 0.1119, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 64.44444444444444, | |
| "grad_norm": 3.1600425243377686, | |
| "learning_rate": 3.2000000000000003e-06, | |
| "loss": 0.1079, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 66.66666666666667, | |
| "grad_norm": 2.9095730781555176, | |
| "learning_rate": 2.9142857142857146e-06, | |
| "loss": 0.1047, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 66.66666666666667, | |
| "eval_loss": 0.7073546648025513, | |
| "eval_runtime": 22.6057, | |
| "eval_samples_per_second": 3.76, | |
| "eval_steps_per_second": 0.088, | |
| "eval_wer": 0.972833365219055, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 68.88888888888889, | |
| "grad_norm": 3.2166309356689453, | |
| "learning_rate": 2.6285714285714286e-06, | |
| "loss": 0.1119, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 71.11111111111111, | |
| "grad_norm": 4.965418815612793, | |
| "learning_rate": 2.342857142857143e-06, | |
| "loss": 0.1101, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 73.33333333333333, | |
| "grad_norm": 2.6848812103271484, | |
| "learning_rate": 2.0571428571428573e-06, | |
| "loss": 0.0907, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 75.55555555555556, | |
| "grad_norm": 2.8126628398895264, | |
| "learning_rate": 1.7714285714285714e-06, | |
| "loss": 0.0854, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 77.77777777777777, | |
| "grad_norm": 2.887021780014038, | |
| "learning_rate": 1.4857142857142858e-06, | |
| "loss": 0.0819, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 77.77777777777777, | |
| "eval_loss": 0.6702935099601746, | |
| "eval_runtime": 20.1033, | |
| "eval_samples_per_second": 4.228, | |
| "eval_steps_per_second": 0.099, | |
| "eval_wer": 0.9627518053971874, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 80.0, | |
| "grad_norm": 2.5661635398864746, | |
| "learning_rate": 1.2000000000000002e-06, | |
| "loss": 0.0825, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 82.22222222222223, | |
| "grad_norm": 2.133279800415039, | |
| "learning_rate": 9.142857142857144e-07, | |
| "loss": 0.0783, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 84.44444444444444, | |
| "grad_norm": 1.9323846101760864, | |
| "learning_rate": 6.285714285714287e-07, | |
| "loss": 0.0767, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 86.66666666666667, | |
| "grad_norm": 1.892685890197754, | |
| "learning_rate": 3.428571428571429e-07, | |
| "loss": 0.0758, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 88.88888888888889, | |
| "grad_norm": 1.9149274826049805, | |
| "learning_rate": 5.714285714285715e-08, | |
| "loss": 0.0728, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 88.88888888888889, | |
| "eval_loss": 0.6913564801216125, | |
| "eval_runtime": 20.0702, | |
| "eval_samples_per_second": 4.235, | |
| "eval_steps_per_second": 0.1, | |
| "eval_wer": 0.9502835538752363, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 88.88888888888889, | |
| "step": 400, | |
| "total_flos": 1.846946575637545e+19, | |
| "train_loss": 0.28228739023208615, | |
| "train_runtime": 3388.702, | |
| "train_samples_per_second": 20.834, | |
| "train_steps_per_second": 0.118 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 400, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 100, | |
| "save_steps": 50, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 1.846946575637545e+19, | |
| "train_batch_size": 40, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |