| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9261208166701747, | |
| "eval_steps": 100, | |
| "global_step": 2200, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.04, | |
| "grad_norm": 2.6430506706237793, | |
| "learning_rate": 9.5e-06, | |
| "loss": 5.6141, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "eval_cer": 0.9770320871299519, | |
| "eval_loss": 2.355087995529175, | |
| "eval_runtime": 329.8418, | |
| "eval_samples_per_second": 28.735, | |
| "eval_steps_per_second": 3.593, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "grad_norm": 2.838773012161255, | |
| "learning_rate": 1.9500000000000003e-05, | |
| "loss": 2.1498, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "eval_cer": 0.6385162879824802, | |
| "eval_loss": 1.961960792541504, | |
| "eval_runtime": 334.0577, | |
| "eval_samples_per_second": 28.372, | |
| "eval_steps_per_second": 3.547, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "grad_norm": 3.848980665206909, | |
| "learning_rate": 2.95e-05, | |
| "loss": 2.6461, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "eval_cer": 0.5852479371162645, | |
| "eval_loss": 2.6564223766326904, | |
| "eval_runtime": 350.7705, | |
| "eval_samples_per_second": 27.021, | |
| "eval_steps_per_second": 3.378, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "grad_norm": 3.9299519062042236, | |
| "learning_rate": 3.9500000000000005e-05, | |
| "loss": 2.359, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "eval_cer": 0.5907913261116108, | |
| "eval_loss": 2.6195201873779297, | |
| "eval_runtime": 341.3159, | |
| "eval_samples_per_second": 27.769, | |
| "eval_steps_per_second": 3.472, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "grad_norm": 4.236645698547363, | |
| "learning_rate": 4.9500000000000004e-05, | |
| "loss": 4.2363, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "eval_cer": 0.624139650385202, | |
| "eval_loss": 2.457566261291504, | |
| "eval_runtime": 356.3471, | |
| "eval_samples_per_second": 26.598, | |
| "eval_steps_per_second": 3.325, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "grad_norm": 13.769584655761719, | |
| "learning_rate": 4.9932862190812725e-05, | |
| "loss": 2.6933, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "eval_cer": 0.6483369833013961, | |
| "eval_loss": 3.4720070362091064, | |
| "eval_runtime": 345.5879, | |
| "eval_samples_per_second": 27.426, | |
| "eval_steps_per_second": 3.429, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "grad_norm": 4.292990207672119, | |
| "learning_rate": 4.986289752650177e-05, | |
| "loss": 4.1075, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "eval_cer": 0.6181098705564898, | |
| "eval_loss": 3.3849761486053467, | |
| "eval_runtime": 353.3198, | |
| "eval_samples_per_second": 26.826, | |
| "eval_steps_per_second": 3.354, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "grad_norm": 4.32611083984375, | |
| "learning_rate": 4.97922261484099e-05, | |
| "loss": 2.6907, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "eval_cer": 0.6191510891243988, | |
| "eval_loss": 2.5412750244140625, | |
| "eval_runtime": 343.4967, | |
| "eval_samples_per_second": 27.593, | |
| "eval_steps_per_second": 3.45, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "grad_norm": 3.173403263092041, | |
| "learning_rate": 4.972155477031802e-05, | |
| "loss": 4.2294, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "eval_cer": 0.5840918423213797, | |
| "eval_loss": 2.6817102432250977, | |
| "eval_runtime": 354.5341, | |
| "eval_samples_per_second": 26.734, | |
| "eval_steps_per_second": 3.342, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "grad_norm": 4.8158440589904785, | |
| "learning_rate": 4.965088339222615e-05, | |
| "loss": 2.2949, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "eval_cer": 0.5736649915920379, | |
| "eval_loss": 1.9513429403305054, | |
| "eval_runtime": 345.1224, | |
| "eval_samples_per_second": 27.463, | |
| "eval_steps_per_second": 3.434, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "grad_norm": 3.0650887489318848, | |
| "learning_rate": 4.958021201413428e-05, | |
| "loss": 2.6115, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "eval_cer": 0.5645799929607759, | |
| "eval_loss": 2.8409903049468994, | |
| "eval_runtime": 352.0824, | |
| "eval_samples_per_second": 26.92, | |
| "eval_steps_per_second": 3.366, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "grad_norm": 3.0540664196014404, | |
| "learning_rate": 4.950954063604241e-05, | |
| "loss": 2.3453, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "eval_cer": 0.598292499315631, | |
| "eval_loss": 1.8713030815124512, | |
| "eval_runtime": 344.1368, | |
| "eval_samples_per_second": 27.541, | |
| "eval_steps_per_second": 3.443, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "grad_norm": 4.366485118865967, | |
| "learning_rate": 4.9438869257950535e-05, | |
| "loss": 3.2578, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "eval_cer": 0.5734816784638849, | |
| "eval_loss": 2.000699758529663, | |
| "eval_runtime": 350.6902, | |
| "eval_samples_per_second": 27.027, | |
| "eval_steps_per_second": 3.379, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "grad_norm": 81.67564392089844, | |
| "learning_rate": 4.936819787985866e-05, | |
| "loss": 1.9657, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "eval_cer": 0.6325427241797349, | |
| "eval_loss": 3.403754472732544, | |
| "eval_runtime": 340.0489, | |
| "eval_samples_per_second": 27.872, | |
| "eval_steps_per_second": 3.485, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "grad_norm": 3.837653875350952, | |
| "learning_rate": 4.929752650176679e-05, | |
| "loss": 2.6865, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "eval_cer": 0.5574918853388604, | |
| "eval_loss": 3.5201330184936523, | |
| "eval_runtime": 351.432, | |
| "eval_samples_per_second": 26.97, | |
| "eval_steps_per_second": 3.372, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "grad_norm": 2.9047458171844482, | |
| "learning_rate": 4.922685512367491e-05, | |
| "loss": 2.4105, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "eval_cer": 0.5886624496499941, | |
| "eval_loss": 3.5269789695739746, | |
| "eval_runtime": 347.023, | |
| "eval_samples_per_second": 27.312, | |
| "eval_steps_per_second": 3.415, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "grad_norm": 2.9270007610321045, | |
| "learning_rate": 4.915618374558304e-05, | |
| "loss": 2.418, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "eval_cer": 0.624877791247898, | |
| "eval_loss": 2.498126983642578, | |
| "eval_runtime": 355.9419, | |
| "eval_samples_per_second": 26.628, | |
| "eval_steps_per_second": 3.329, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "grad_norm": 3.458078622817993, | |
| "learning_rate": 4.908551236749117e-05, | |
| "loss": 2.6118, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "eval_cer": 0.5718074185600877, | |
| "eval_loss": 3.407224655151367, | |
| "eval_runtime": 343.8365, | |
| "eval_samples_per_second": 27.565, | |
| "eval_steps_per_second": 3.446, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "grad_norm": 1.9933278560638428, | |
| "learning_rate": 4.90148409893993e-05, | |
| "loss": 3.2046, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "eval_cer": 0.5539844941535333, | |
| "eval_loss": 1.928993821144104, | |
| "eval_runtime": 357.2509, | |
| "eval_samples_per_second": 26.53, | |
| "eval_steps_per_second": 3.317, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "grad_norm": 4.917328357696533, | |
| "learning_rate": 4.8944169611307425e-05, | |
| "loss": 3.1904, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "eval_cer": 0.5568001838019632, | |
| "eval_loss": 2.729193925857544, | |
| "eval_runtime": 341.3463, | |
| "eval_samples_per_second": 27.767, | |
| "eval_steps_per_second": 3.472, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "grad_norm": 2.842353343963623, | |
| "learning_rate": 4.8873498233215547e-05, | |
| "loss": 2.1649, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "eval_cer": 0.544757733369833, | |
| "eval_loss": 2.421653985977173, | |
| "eval_runtime": 358.7261, | |
| "eval_samples_per_second": 26.421, | |
| "eval_steps_per_second": 3.303, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "grad_norm": 4.818225383758545, | |
| "learning_rate": 4.880282685512368e-05, | |
| "loss": 2.2775, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "eval_cer": 0.5533832270931915, | |
| "eval_loss": 2.188101291656494, | |
| "eval_runtime": 354.4215, | |
| "eval_samples_per_second": 26.742, | |
| "eval_steps_per_second": 3.343, | |
| "step": 2200 | |
| } | |
| ], | |
| "logging_steps": 100, | |
| "max_steps": 71250, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 30, | |
| "save_steps": 100, | |
| "total_flos": 2.411832086421037e+19, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |