{ "best_metric": 0.2726672291755676, "best_model_checkpoint": "nrshoudi/wav2vec_arabic_mdd/checkpoint-8185", "epoch": 20.0, "eval_steps": 500, "global_step": 32740, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 1.0, "learning_rate": 5e-05, "loss": 8.1087, "step": 1637 }, { "epoch": 1.0, "eval_loss": 3.1701228618621826, "eval_per": 1.0, "eval_runtime": 115.1384, "eval_samples_per_second": 7.061, "eval_steps_per_second": 3.535, "eval_wer": 1.0, "step": 1637 }, { "epoch": 2.0, "learning_rate": 0.0001, "loss": 1.1869, "step": 3274 }, { "epoch": 2.0, "eval_loss": 0.40448808670043945, "eval_per": 0.06459059643833034, "eval_runtime": 117.2186, "eval_samples_per_second": 6.936, "eval_steps_per_second": 3.472, "eval_wer": 0.07997760253837899, "step": 3274 }, { "epoch": 3.0, "learning_rate": 9.444444444444444e-05, "loss": 0.2, "step": 4911 }, { "epoch": 3.0, "eval_loss": 0.3259519040584564, "eval_per": 0.04635823989446905, "eval_runtime": 116.6876, "eval_samples_per_second": 6.967, "eval_steps_per_second": 3.488, "eval_wer": 0.05907330502543045, "step": 4911 }, { "epoch": 4.0, "learning_rate": 8.888888888888889e-05, "loss": 0.1283, "step": 6548 }, { "epoch": 4.0, "eval_loss": 0.3042123317718506, "eval_per": 0.04748892867238293, "eval_runtime": 116.8704, "eval_samples_per_second": 6.956, "eval_steps_per_second": 3.482, "eval_wer": 0.06182632634968037, "step": 6548 }, { "epoch": 5.0, "learning_rate": 8.333333333333334e-05, "loss": 0.1037, "step": 8185 }, { "epoch": 5.0, "eval_loss": 0.2726672291755676, "eval_per": 0.04101102421558466, "eval_runtime": 117.5633, "eval_samples_per_second": 6.915, "eval_steps_per_second": 3.462, "eval_wer": 0.053147309971536556, "step": 8185 }, { "epoch": 6.0, "learning_rate": 7.777777777777778e-05, "loss": 0.0844, "step": 9822 }, { "epoch": 6.0, "eval_loss": 0.31839433312416077, "eval_per": 0.040916800150758506, "eval_runtime": 117.5398, "eval_samples_per_second": 6.917, "eval_steps_per_second": 3.463, "eval_wer": 0.05431384443096449, "step": 9822 }, { "epoch": 7.0, "learning_rate": 7.222222222222222e-05, "loss": 0.0738, "step": 11459 }, { "epoch": 7.0, "eval_loss": 0.28862160444259644, "eval_per": 0.03655893715254876, "eval_runtime": 117.5297, "eval_samples_per_second": 6.917, "eval_steps_per_second": 3.463, "eval_wer": 0.048527833512201954, "step": 11459 }, { "epoch": 8.0, "learning_rate": 6.666666666666667e-05, "loss": 0.0613, "step": 13096 }, { "epoch": 8.0, "eval_loss": 0.3345281481742859, "eval_per": 0.037359841703571095, "eval_runtime": 117.0501, "eval_samples_per_second": 6.946, "eval_steps_per_second": 3.477, "eval_wer": 0.04880780178246465, "step": 13096 }, { "epoch": 9.0, "learning_rate": 6.111111111111112e-05, "loss": 0.0573, "step": 14733 }, { "epoch": 9.0, "eval_loss": 0.34689363837242126, "eval_per": 0.03943277112974654, "eval_runtime": 117.6593, "eval_samples_per_second": 6.91, "eval_steps_per_second": 3.459, "eval_wer": 0.05053427278241799, "step": 14733 }, { "epoch": 10.0, "learning_rate": 5.555555555555556e-05, "loss": 0.0514, "step": 16370 }, { "epoch": 10.0, "eval_loss": 0.3245474100112915, "eval_per": 0.038608310562517664, "eval_runtime": 117.7221, "eval_samples_per_second": 6.906, "eval_steps_per_second": 3.457, "eval_wer": 0.051047547944566285, "step": 16370 }, { "epoch": 11.0, "learning_rate": 5e-05, "loss": 0.0469, "step": 18007 }, { "epoch": 11.0, "eval_loss": 0.30943331122398376, "eval_per": 0.03738339771977763, "eval_runtime": 118.4378, "eval_samples_per_second": 6.864, "eval_steps_per_second": 3.436, "eval_wer": 0.049181092809481596, "step": 18007 }, { "epoch": 12.0, "learning_rate": 4.4444444444444447e-05, "loss": 0.0375, "step": 19644 }, { "epoch": 12.0, "eval_loss": 0.36555343866348267, "eval_per": 0.03919721096768115, "eval_runtime": 117.4812, "eval_samples_per_second": 6.92, "eval_steps_per_second": 3.464, "eval_wer": 0.05207409826886286, "step": 19644 }, { "epoch": 13.0, "learning_rate": 3.888888888888889e-05, "loss": 0.0356, "step": 21281 }, { "epoch": 13.0, "eval_loss": 0.3295977711677551, "eval_per": 0.03561669650428719, "eval_runtime": 120.3969, "eval_samples_per_second": 6.753, "eval_steps_per_second": 3.38, "eval_wer": 0.04722131491764267, "step": 21281 }, { "epoch": 14.0, "learning_rate": 3.3333333333333335e-05, "loss": 0.0291, "step": 22918 }, { "epoch": 14.0, "eval_loss": 0.3301165699958801, "eval_per": 0.033637991142937904, "eval_runtime": 120.2853, "eval_samples_per_second": 6.759, "eval_steps_per_second": 3.384, "eval_wer": 0.04479492324203257, "step": 22918 }, { "epoch": 15.0, "learning_rate": 2.777777777777778e-05, "loss": 0.0242, "step": 24555 }, { "epoch": 15.0, "eval_loss": 0.3575274646282196, "eval_per": 0.03573447658531989, "eval_runtime": 120.2073, "eval_samples_per_second": 6.763, "eval_steps_per_second": 3.386, "eval_wer": 0.046008119079837616, "step": 24555 }, { "epoch": 16.0, "learning_rate": 2.2222222222222223e-05, "loss": 0.0216, "step": 26192 }, { "epoch": 16.0, "eval_loss": 0.3376178741455078, "eval_per": 0.033496655045698674, "eval_runtime": 119.4419, "eval_samples_per_second": 6.807, "eval_steps_per_second": 3.408, "eval_wer": 0.04428164807988428, "step": 26192 }, { "epoch": 17.0, "learning_rate": 1.6666666666666667e-05, "loss": 0.0208, "step": 27829 }, { "epoch": 17.0, "eval_loss": 0.3688383400440216, "eval_per": 0.033190426835013664, "eval_runtime": 118.4061, "eval_samples_per_second": 6.866, "eval_steps_per_second": 3.437, "eval_wer": 0.04358172740422752, "step": 27829 }, { "epoch": 18.0, "learning_rate": 1.1111111111111112e-05, "loss": 0.018, "step": 29466 }, { "epoch": 18.0, "eval_loss": 0.3673172891139984, "eval_per": 0.03403844341844907, "eval_runtime": 118.1876, "eval_samples_per_second": 6.879, "eval_steps_per_second": 3.444, "eval_wer": 0.04446829359339275, "step": 29466 }, { "epoch": 19.0, "learning_rate": 5.555555555555556e-06, "loss": 0.0164, "step": 31103 }, { "epoch": 19.0, "eval_loss": 0.35759395360946655, "eval_per": 0.03271930651088288, "eval_runtime": 119.0412, "eval_samples_per_second": 6.83, "eval_steps_per_second": 3.419, "eval_wer": 0.043208436377210585, "step": 31103 }, { "epoch": 20.0, "learning_rate": 0.0, "loss": 0.0128, "step": 32740 }, { "epoch": 20.0, "eval_loss": 0.3501162827014923, "eval_per": 0.03269575049467634, "eval_runtime": 118.2142, "eval_samples_per_second": 6.877, "eval_steps_per_second": 3.443, "eval_wer": 0.04316177499883347, "step": 32740 }, { "epoch": 20.0, "step": 32740, "total_flos": 1.0473975577487264e+19, "train_loss": 0.5159456469186866, "train_runtime": 19614.1198, "train_samples_per_second": 3.337, "train_steps_per_second": 1.669 } ], "logging_steps": 500, "max_steps": 32740, "num_train_epochs": 20, "save_steps": 500, "total_flos": 1.0473975577487264e+19, "trial_name": null, "trial_params": null }