{ "best_metric": null, "best_model_checkpoint": null, "epoch": 4.602431690417048, "global_step": 17600, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.21, "learning_rate": 0.0002976391096979332, "loss": 6.4525, "step": 800 }, { "epoch": 0.21, "eval_loss": 1.4264159202575684, "eval_runtime": 579.6055, "eval_samples_per_second": 37.262, "eval_steps_per_second": 4.658, "eval_wer": 0.9522086267891338, "step": 800 }, { "epoch": 0.42, "learning_rate": 0.00029127980922098566, "loss": 1.105, "step": 1600 }, { "epoch": 0.42, "eval_loss": 0.5966897010803223, "eval_runtime": 638.739, "eval_samples_per_second": 33.812, "eval_steps_per_second": 4.227, "eval_wer": 0.6514394209860115, "step": 1600 }, { "epoch": 0.63, "learning_rate": 0.00028492050874403814, "loss": 0.7933, "step": 2400 }, { "epoch": 0.63, "eval_loss": 0.44666677713394165, "eval_runtime": 578.27, "eval_samples_per_second": 37.348, "eval_steps_per_second": 4.669, "eval_wer": 0.5045276037778715, "step": 2400 }, { "epoch": 0.84, "learning_rate": 0.00027856120826709056, "loss": 0.6926, "step": 3200 }, { "epoch": 0.84, "eval_loss": 0.3889370858669281, "eval_runtime": 574.1787, "eval_samples_per_second": 37.614, "eval_steps_per_second": 4.702, "eval_wer": 0.4508941611762033, "step": 3200 }, { "epoch": 1.05, "learning_rate": 0.0002722019077901431, "loss": 0.6258, "step": 4000 }, { "epoch": 1.05, "eval_loss": 0.33456486463546753, "eval_runtime": 575.0581, "eval_samples_per_second": 37.556, "eval_steps_per_second": 4.695, "eval_wer": 0.39778001363149523, "step": 4000 }, { "epoch": 1.26, "learning_rate": 0.0002658426073131955, "loss": 0.5666, "step": 4800 }, { "epoch": 1.26, "eval_loss": 0.31256240606307983, "eval_runtime": 592.8115, "eval_samples_per_second": 36.431, "eval_steps_per_second": 4.555, "eval_wer": 0.3761481289149979, "step": 4800 }, { "epoch": 1.46, "learning_rate": 0.000259483306836248, "loss": 0.5501, "step": 5600 }, { "epoch": 1.46, "eval_loss": 0.2980588376522064, "eval_runtime": 662.1285, "eval_samples_per_second": 32.618, "eval_steps_per_second": 4.078, "eval_wer": 0.3593684073869722, "step": 5600 }, { "epoch": 1.67, "learning_rate": 0.00025313195548489663, "loss": 0.5275, "step": 6400 }, { "epoch": 1.67, "eval_loss": 0.28137603402137756, "eval_runtime": 588.9138, "eval_samples_per_second": 36.673, "eval_steps_per_second": 4.585, "eval_wer": 0.3457206841712375, "step": 6400 }, { "epoch": 1.88, "learning_rate": 0.0002467726550079491, "loss": 0.5203, "step": 7200 }, { "epoch": 1.88, "eval_loss": 0.2654956877231598, "eval_runtime": 586.4688, "eval_samples_per_second": 36.825, "eval_steps_per_second": 4.604, "eval_wer": 0.3303041121677323, "step": 7200 }, { "epoch": 2.09, "learning_rate": 0.00024041335453100157, "loss": 0.4842, "step": 8000 }, { "epoch": 2.09, "eval_loss": 0.2574324607849121, "eval_runtime": 641.5915, "eval_samples_per_second": 33.662, "eval_steps_per_second": 4.208, "eval_wer": 0.31566648274966735, "step": 8000 }, { "epoch": 2.3, "learning_rate": 0.00023405405405405404, "loss": 0.4542, "step": 8800 }, { "epoch": 2.3, "eval_loss": 0.2577317953109741, "eval_runtime": 583.455, "eval_samples_per_second": 37.016, "eval_steps_per_second": 4.628, "eval_wer": 0.31417350946090683, "step": 8800 }, { "epoch": 2.51, "learning_rate": 0.0002276947535771065, "loss": 0.4431, "step": 9600 }, { "epoch": 2.51, "eval_loss": 0.23904646933078766, "eval_runtime": 581.9022, "eval_samples_per_second": 37.114, "eval_steps_per_second": 4.64, "eval_wer": 0.2938236344162799, "step": 9600 }, { "epoch": 2.72, "learning_rate": 0.00022133545310015898, "loss": 0.4235, "step": 10400 }, { "epoch": 2.72, "eval_loss": 0.24016053974628448, "eval_runtime": 582.7306, "eval_samples_per_second": 37.062, "eval_steps_per_second": 4.633, "eval_wer": 0.2884684041413781, "step": 10400 }, { "epoch": 2.93, "learning_rate": 0.0002149920508744038, "loss": 0.4202, "step": 11200 }, { "epoch": 2.93, "eval_loss": 0.22816315293312073, "eval_runtime": 643.2534, "eval_samples_per_second": 33.575, "eval_steps_per_second": 4.197, "eval_wer": 0.2846710590373568, "step": 11200 }, { "epoch": 3.14, "learning_rate": 0.00020863275039745627, "loss": 0.4018, "step": 12000 }, { "epoch": 3.14, "eval_loss": 0.22759725153446198, "eval_runtime": 584.1242, "eval_samples_per_second": 36.973, "eval_steps_per_second": 4.622, "eval_wer": 0.27702768491772417, "step": 12000 }, { "epoch": 3.35, "learning_rate": 0.00020227344992050872, "loss": 0.3986, "step": 12800 }, { "epoch": 3.35, "eval_loss": 0.224422425031662, "eval_runtime": 582.249, "eval_samples_per_second": 37.092, "eval_steps_per_second": 4.637, "eval_wer": 0.27022816526565185, "step": 12800 }, { "epoch": 3.56, "learning_rate": 0.00019591414944356117, "loss": 0.3792, "step": 13600 }, { "epoch": 3.56, "eval_loss": 0.21611596643924713, "eval_runtime": 589.0669, "eval_samples_per_second": 36.663, "eval_steps_per_second": 4.584, "eval_wer": 0.2742527019570932, "step": 13600 }, { "epoch": 3.77, "learning_rate": 0.00018955484896661365, "loss": 0.3736, "step": 14400 }, { "epoch": 3.77, "eval_loss": 0.20533400774002075, "eval_runtime": 644.3789, "eval_samples_per_second": 33.516, "eval_steps_per_second": 4.19, "eval_wer": 0.2546979974684366, "step": 14400 }, { "epoch": 3.97, "learning_rate": 0.0001831955484896661, "loss": 0.37, "step": 15200 }, { "epoch": 3.97, "eval_loss": 0.19958257675170898, "eval_runtime": 585.9573, "eval_samples_per_second": 36.858, "eval_steps_per_second": 4.608, "eval_wer": 0.25521729252539677, "step": 15200 }, { "epoch": 4.18, "learning_rate": 0.00017683624801271858, "loss": 0.3499, "step": 16000 }, { "epoch": 4.18, "eval_loss": 0.19865886867046356, "eval_runtime": 587.8211, "eval_samples_per_second": 36.741, "eval_steps_per_second": 4.593, "eval_wer": 0.24392262503651294, "step": 16000 }, { "epoch": 4.39, "learning_rate": 0.00017048489666136724, "loss": 0.3416, "step": 16800 }, { "epoch": 4.39, "eval_loss": 0.1961081475019455, "eval_runtime": 586.8214, "eval_samples_per_second": 36.803, "eval_steps_per_second": 4.601, "eval_wer": 0.24739541072993412, "step": 16800 }, { "epoch": 4.6, "learning_rate": 0.0001641255961844197, "loss": 0.3329, "step": 17600 }, { "epoch": 4.6, "eval_loss": 0.19132652878761292, "eval_runtime": 586.4075, "eval_samples_per_second": 36.829, "eval_steps_per_second": 4.604, "eval_wer": 0.2405309791957418, "step": 17600 } ], "max_steps": 38240, "num_train_epochs": 10, "total_flos": 5.592197758263177e+19, "trial_name": null, "trial_params": null }