{ "best_metric": 0.1885096763522082, "best_model_checkpoint": "./en-xlsr/checkpoint-6000", "epoch": 27.906976744186046, "eval_steps": 600, "global_step": 6000, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.47, "learning_rate": 0.0002983555207517619, "loss": 4.6031, "step": 100 }, { "epoch": 0.93, "learning_rate": 0.0002936570086139389, "loss": 2.9545, "step": 200 }, { "epoch": 1.4, "learning_rate": 0.0002889584964761159, "loss": 2.8375, "step": 300 }, { "epoch": 1.86, "learning_rate": 0.00028425998433829287, "loss": 1.6291, "step": 400 }, { "epoch": 2.33, "learning_rate": 0.0002795614722004698, "loss": 0.8607, "step": 500 }, { "epoch": 2.79, "learning_rate": 0.0002748629600626468, "loss": 0.7055, "step": 600 }, { "epoch": 2.79, "eval_cer": 0.1303822520884634, "eval_loss": 0.4910881221294403, "eval_runtime": 23.4039, "eval_samples_per_second": 74.261, "eval_steps_per_second": 4.657, "eval_wer": 0.3307603241991509, "step": 600 }, { "epoch": 3.26, "learning_rate": 0.0002701644479248238, "loss": 0.5932, "step": 700 }, { "epoch": 3.72, "learning_rate": 0.0002654659357870008, "loss": 0.5297, "step": 800 }, { "epoch": 4.19, "learning_rate": 0.00026076742364917777, "loss": 0.4983, "step": 900 }, { "epoch": 4.65, "learning_rate": 0.0002560689115113547, "loss": 0.4337, "step": 1000 }, { "epoch": 5.12, "learning_rate": 0.0002513703993735317, "loss": 0.4183, "step": 1100 }, { "epoch": 5.58, "learning_rate": 0.0002466718872357087, "loss": 0.3761, "step": 1200 }, { "epoch": 5.58, "eval_cer": 0.10525165120935263, "eval_loss": 0.3983539938926697, "eval_runtime": 23.2773, "eval_samples_per_second": 74.665, "eval_steps_per_second": 4.683, "eval_wer": 0.2532943706235871, "step": 1200 }, { "epoch": 6.05, "learning_rate": 0.00024197337509788565, "loss": 0.3694, "step": 1300 }, { "epoch": 6.51, "learning_rate": 0.00023727486296006264, "loss": 0.3278, "step": 1400 }, { "epoch": 6.98, "learning_rate": 0.0002325763508222396, "loss": 0.3282, "step": 1500 }, { "epoch": 7.44, "learning_rate": 0.0002278778386844166, "loss": 0.2941, "step": 1600 }, { "epoch": 7.91, "learning_rate": 0.00022317932654659358, "loss": 0.2994, "step": 1700 }, { "epoch": 8.37, "learning_rate": 0.00021848081440877052, "loss": 0.278, "step": 1800 }, { "epoch": 8.37, "eval_cer": 0.10237497986330057, "eval_loss": 0.4070082902908325, "eval_runtime": 22.7099, "eval_samples_per_second": 76.531, "eval_steps_per_second": 4.8, "eval_wer": 0.24452776093069417, "step": 1800 }, { "epoch": 8.84, "learning_rate": 0.0002137823022709475, "loss": 0.2658, "step": 1900 }, { "epoch": 9.3, "learning_rate": 0.00020908379013312447, "loss": 0.2617, "step": 2000 }, { "epoch": 9.77, "learning_rate": 0.00020438527799530146, "loss": 0.2484, "step": 2100 }, { "epoch": 10.23, "learning_rate": 0.00019968676585747845, "loss": 0.2402, "step": 2200 }, { "epoch": 10.7, "learning_rate": 0.00019498825371965542, "loss": 0.2326, "step": 2300 }, { "epoch": 11.16, "learning_rate": 0.0001902897415818324, "loss": 0.2196, "step": 2400 }, { "epoch": 11.16, "eval_cer": 0.09736957172117, "eval_loss": 0.4033309519290924, "eval_runtime": 23.1418, "eval_samples_per_second": 75.102, "eval_steps_per_second": 4.71, "eval_wer": 0.22434801786403485, "step": 2400 }, { "epoch": 11.63, "learning_rate": 0.0001855912294440094, "loss": 0.2127, "step": 2500 }, { "epoch": 12.09, "learning_rate": 0.00018089271730618636, "loss": 0.2099, "step": 2600 }, { "epoch": 12.56, "learning_rate": 0.00017619420516836332, "loss": 0.1917, "step": 2700 }, { "epoch": 13.02, "learning_rate": 0.00017149569303054031, "loss": 0.1927, "step": 2800 }, { "epoch": 13.49, "learning_rate": 0.00016679718089271728, "loss": 0.1794, "step": 2900 }, { "epoch": 13.95, "learning_rate": 0.00016209866875489427, "loss": 0.1842, "step": 3000 }, { "epoch": 13.95, "eval_cer": 0.0927553908821025, "eval_loss": 0.4269569218158722, "eval_runtime": 22.8362, "eval_samples_per_second": 76.107, "eval_steps_per_second": 4.773, "eval_wer": 0.21061917626950433, "step": 3000 }, { "epoch": 14.42, "learning_rate": 0.00015740015661707126, "loss": 0.1719, "step": 3100 }, { "epoch": 14.88, "learning_rate": 0.00015270164447924822, "loss": 0.1652, "step": 3200 }, { "epoch": 15.35, "learning_rate": 0.00014800313234142518, "loss": 0.1617, "step": 3300 }, { "epoch": 15.81, "learning_rate": 0.00014330462020360218, "loss": 0.1572, "step": 3400 }, { "epoch": 16.28, "learning_rate": 0.00013860610806577917, "loss": 0.1493, "step": 3500 }, { "epoch": 16.74, "learning_rate": 0.00013390759592795613, "loss": 0.1533, "step": 3600 }, { "epoch": 16.74, "eval_cer": 0.09155869560214484, "eval_loss": 0.45823100209236145, "eval_runtime": 23.9029, "eval_samples_per_second": 72.711, "eval_steps_per_second": 4.56, "eval_wer": 0.20714561393835806, "step": 3600 }, { "epoch": 17.21, "learning_rate": 0.00012920908379013312, "loss": 0.1469, "step": 3700 }, { "epoch": 17.67, "learning_rate": 0.00012451057165231008, "loss": 0.1389, "step": 3800 }, { "epoch": 18.14, "learning_rate": 0.00011981205951448707, "loss": 0.1307, "step": 3900 }, { "epoch": 18.6, "learning_rate": 0.00011511354737666405, "loss": 0.1338, "step": 4000 }, { "epoch": 19.07, "learning_rate": 0.00011041503523884103, "loss": 0.1242, "step": 4100 }, { "epoch": 19.53, "learning_rate": 0.000105716523101018, "loss": 0.1257, "step": 4200 }, { "epoch": 19.53, "eval_cer": 0.09014337329988724, "eval_loss": 0.4684600532054901, "eval_runtime": 23.1553, "eval_samples_per_second": 75.058, "eval_steps_per_second": 4.707, "eval_wer": 0.2001433533660473, "step": 4200 }, { "epoch": 20.0, "learning_rate": 0.00010101801096319498, "loss": 0.1259, "step": 4300 }, { "epoch": 20.47, "learning_rate": 9.631949882537196e-05, "loss": 0.1155, "step": 4400 }, { "epoch": 20.93, "learning_rate": 9.162098668754895e-05, "loss": 0.1186, "step": 4500 }, { "epoch": 21.4, "learning_rate": 8.692247454972591e-05, "loss": 0.1078, "step": 4600 }, { "epoch": 21.86, "learning_rate": 8.222396241190289e-05, "loss": 0.1128, "step": 4700 }, { "epoch": 22.33, "learning_rate": 7.752545027407988e-05, "loss": 0.1071, "step": 4800 }, { "epoch": 22.33, "eval_cer": 0.08777299611074034, "eval_loss": 0.5087743997573853, "eval_runtime": 22.8722, "eval_samples_per_second": 75.987, "eval_steps_per_second": 4.766, "eval_wer": 0.19650438330484646, "step": 4800 }, { "epoch": 22.79, "learning_rate": 7.282693813625684e-05, "loss": 0.1049, "step": 4900 }, { "epoch": 23.26, "learning_rate": 6.812842599843383e-05, "loss": 0.0993, "step": 5000 }, { "epoch": 23.72, "learning_rate": 6.34299138606108e-05, "loss": 0.0983, "step": 5100 }, { "epoch": 24.19, "learning_rate": 5.873140172278778e-05, "loss": 0.0949, "step": 5200 }, { "epoch": 24.65, "learning_rate": 5.4032889584964754e-05, "loss": 0.0914, "step": 5300 }, { "epoch": 25.12, "learning_rate": 4.933437744714174e-05, "loss": 0.0967, "step": 5400 }, { "epoch": 25.12, "eval_cer": 0.08720916852691414, "eval_loss": 0.5224480032920837, "eval_runtime": 23.2681, "eval_samples_per_second": 74.694, "eval_steps_per_second": 4.685, "eval_wer": 0.19126647185311793, "step": 5400 }, { "epoch": 25.58, "learning_rate": 4.463586530931871e-05, "loss": 0.091, "step": 5500 }, { "epoch": 26.05, "learning_rate": 3.993735317149569e-05, "loss": 0.0866, "step": 5600 }, { "epoch": 26.51, "learning_rate": 3.523884103367267e-05, "loss": 0.0908, "step": 5700 }, { "epoch": 26.98, "learning_rate": 3.0540328895849644e-05, "loss": 0.0855, "step": 5800 }, { "epoch": 27.44, "learning_rate": 2.584181675802662e-05, "loss": 0.0869, "step": 5900 }, { "epoch": 27.91, "learning_rate": 2.1143304620203598e-05, "loss": 0.0839, "step": 6000 }, { "epoch": 27.91, "eval_cer": 0.08595493982003544, "eval_loss": 0.5378865599632263, "eval_runtime": 23.1463, "eval_samples_per_second": 75.088, "eval_steps_per_second": 4.709, "eval_wer": 0.1885096763522082, "step": 6000 } ], "logging_steps": 100, "max_steps": 6450, "num_train_epochs": 30, "save_steps": 600, "total_flos": 4.161559358313857e+19, "trial_name": null, "trial_params": null }