| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 2.9888444537992003, | |
| "eval_steps": 100, | |
| "global_step": 7100, | |
| "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 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "grad_norm": 2.9436004161834717, | |
| "learning_rate": 4.87321554770318e-05, | |
| "loss": 1.6308, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "eval_cer": 0.5162586523796489, | |
| "eval_loss": 3.3528623580932617, | |
| "eval_runtime": 360.9925, | |
| "eval_samples_per_second": 26.255, | |
| "eval_steps_per_second": 3.283, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 1.01, | |
| "grad_norm": 7.935622692108154, | |
| "learning_rate": 4.866148409893993e-05, | |
| "loss": 2.4462, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 1.01, | |
| "eval_cer": 0.575390579171718, | |
| "eval_loss": 2.0690765380859375, | |
| "eval_runtime": 343.7001, | |
| "eval_samples_per_second": 27.576, | |
| "eval_steps_per_second": 3.448, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 1.05, | |
| "grad_norm": 2.40964937210083, | |
| "learning_rate": 4.859081272084806e-05, | |
| "loss": 2.3368, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.05, | |
| "eval_cer": 0.5851746118650033, | |
| "eval_loss": 1.9030027389526367, | |
| "eval_runtime": 352.7743, | |
| "eval_samples_per_second": 26.867, | |
| "eval_steps_per_second": 3.359, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.09, | |
| "grad_norm": 3.7618064880371094, | |
| "learning_rate": 4.852084805653711e-05, | |
| "loss": 2.4157, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 1.09, | |
| "eval_cer": 0.5566339798991045, | |
| "eval_loss": 3.213930368423462, | |
| "eval_runtime": 341.3273, | |
| "eval_samples_per_second": 27.768, | |
| "eval_steps_per_second": 3.472, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 1.14, | |
| "grad_norm": 9.61824893951416, | |
| "learning_rate": 4.845017667844523e-05, | |
| "loss": 2.7299, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 1.14, | |
| "eval_cer": 0.5819702983848891, | |
| "eval_loss": 1.7893775701522827, | |
| "eval_runtime": 353.5497, | |
| "eval_samples_per_second": 26.808, | |
| "eval_steps_per_second": 3.352, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 1.18, | |
| "grad_norm": 2.6824772357940674, | |
| "learning_rate": 4.8379505300353364e-05, | |
| "loss": 2.934, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 1.18, | |
| "eval_cer": 0.5119813460560791, | |
| "eval_loss": 2.804206609725952, | |
| "eval_runtime": 340.4666, | |
| "eval_samples_per_second": 27.838, | |
| "eval_steps_per_second": 3.481, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 1.22, | |
| "grad_norm": 4.691165924072266, | |
| "learning_rate": 4.8308833922261485e-05, | |
| "loss": 2.0012, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 1.22, | |
| "eval_cer": 0.5138047006374409, | |
| "eval_loss": 3.412604331970215, | |
| "eval_runtime": 349.7896, | |
| "eval_samples_per_second": 27.096, | |
| "eval_steps_per_second": 3.388, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 1.26, | |
| "grad_norm": 2.8883514404296875, | |
| "learning_rate": 4.823816254416961e-05, | |
| "loss": 3.5496, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.26, | |
| "eval_cer": 0.5417074029173673, | |
| "eval_loss": 1.7423540353775024, | |
| "eval_runtime": 339.3199, | |
| "eval_samples_per_second": 27.932, | |
| "eval_steps_per_second": 3.492, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.3, | |
| "grad_norm": 3.9499669075012207, | |
| "learning_rate": 4.816749116607774e-05, | |
| "loss": 2.4353, | |
| "step": 3100 | |
| }, | |
| { | |
| "epoch": 1.3, | |
| "eval_cer": 0.5479229400492746, | |
| "eval_loss": 3.4527170658111572, | |
| "eval_runtime": 354.1425, | |
| "eval_samples_per_second": 26.763, | |
| "eval_steps_per_second": 3.346, | |
| "step": 3100 | |
| }, | |
| { | |
| "epoch": 1.35, | |
| "grad_norm": 2.9536569118499756, | |
| "learning_rate": 4.809681978798587e-05, | |
| "loss": 2.9787, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 1.35, | |
| "eval_cer": 0.5677500879903015, | |
| "eval_loss": 2.6674227714538574, | |
| "eval_runtime": 344.0522, | |
| "eval_samples_per_second": 27.548, | |
| "eval_steps_per_second": 3.444, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 1.39, | |
| "grad_norm": 3.094930648803711, | |
| "learning_rate": 4.8026148409894e-05, | |
| "loss": 2.2166, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 1.39, | |
| "eval_cer": 0.5857514371749247, | |
| "eval_loss": 3.2282423973083496, | |
| "eval_runtime": 358.5567, | |
| "eval_samples_per_second": 26.434, | |
| "eval_steps_per_second": 3.305, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 1.43, | |
| "grad_norm": 5.626856803894043, | |
| "learning_rate": 4.795547703180212e-05, | |
| "loss": 2.6222, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 1.43, | |
| "eval_cer": 0.5495043213014743, | |
| "eval_loss": 1.7686785459518433, | |
| "eval_runtime": 342.362, | |
| "eval_samples_per_second": 27.684, | |
| "eval_steps_per_second": 3.461, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 1.47, | |
| "grad_norm": 5.24758768081665, | |
| "learning_rate": 4.7884805653710253e-05, | |
| "loss": 2.078, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 1.47, | |
| "eval_cer": 0.5179402448085723, | |
| "eval_loss": 2.244272470474243, | |
| "eval_runtime": 354.1685, | |
| "eval_samples_per_second": 26.761, | |
| "eval_steps_per_second": 3.346, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 1.52, | |
| "grad_norm": 6.822329044342041, | |
| "learning_rate": 4.7814134275618375e-05, | |
| "loss": 2.7694, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 1.52, | |
| "eval_cer": 0.5457598451370693, | |
| "eval_loss": 2.137202501296997, | |
| "eval_runtime": 345.4818, | |
| "eval_samples_per_second": 27.434, | |
| "eval_steps_per_second": 3.43, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 1.56, | |
| "grad_norm": 2.648332357406616, | |
| "learning_rate": 4.77434628975265e-05, | |
| "loss": 1.74, | |
| "step": 3700 | |
| }, | |
| { | |
| "epoch": 1.56, | |
| "eval_cer": 0.5568441789527199, | |
| "eval_loss": 3.1145806312561035, | |
| "eval_runtime": 362.0538, | |
| "eval_samples_per_second": 26.178, | |
| "eval_steps_per_second": 3.273, | |
| "step": 3700 | |
| }, | |
| { | |
| "epoch": 1.6, | |
| "grad_norm": 17.834945678710938, | |
| "learning_rate": 4.767279151943463e-05, | |
| "loss": 1.7586, | |
| "step": 3800 | |
| }, | |
| { | |
| "epoch": 1.6, | |
| "eval_cer": 0.5344653122678034, | |
| "eval_loss": 1.8897360563278198, | |
| "eval_runtime": 345.2891, | |
| "eval_samples_per_second": 27.449, | |
| "eval_steps_per_second": 3.432, | |
| "step": 3800 | |
| }, | |
| { | |
| "epoch": 1.64, | |
| "grad_norm": 7.307125568389893, | |
| "learning_rate": 4.760212014134276e-05, | |
| "loss": 1.6227, | |
| "step": 3900 | |
| }, | |
| { | |
| "epoch": 1.64, | |
| "eval_cer": 0.551083258378632, | |
| "eval_loss": 1.9913257360458374, | |
| "eval_runtime": 354.7374, | |
| "eval_samples_per_second": 26.718, | |
| "eval_steps_per_second": 3.34, | |
| "step": 3900 | |
| }, | |
| { | |
| "epoch": 1.68, | |
| "grad_norm": 4.344304084777832, | |
| "learning_rate": 4.753144876325089e-05, | |
| "loss": 3.1757, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.68, | |
| "eval_cer": 0.5003299636306754, | |
| "eval_loss": 1.8998807668685913, | |
| "eval_runtime": 348.4123, | |
| "eval_samples_per_second": 27.203, | |
| "eval_steps_per_second": 3.401, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.73, | |
| "grad_norm": 2.8654837608337402, | |
| "learning_rate": 4.746077738515901e-05, | |
| "loss": 2.226, | |
| "step": 4100 | |
| }, | |
| { | |
| "epoch": 1.73, | |
| "eval_cer": 0.5324904188338352, | |
| "eval_loss": 3.042088031768799, | |
| "eval_runtime": 354.1261, | |
| "eval_samples_per_second": 26.764, | |
| "eval_steps_per_second": 3.346, | |
| "step": 4100 | |
| }, | |
| { | |
| "epoch": 1.77, | |
| "grad_norm": 7.595367908477783, | |
| "learning_rate": 4.7390106007067143e-05, | |
| "loss": 2.7925, | |
| "step": 4200 | |
| }, | |
| { | |
| "epoch": 1.77, | |
| "eval_cer": 0.5470479253842243, | |
| "eval_loss": 3.0831844806671143, | |
| "eval_runtime": 344.2619, | |
| "eval_samples_per_second": 27.531, | |
| "eval_steps_per_second": 3.442, | |
| "step": 4200 | |
| }, | |
| { | |
| "epoch": 1.81, | |
| "grad_norm": 7.425869941711426, | |
| "learning_rate": 4.7319434628975265e-05, | |
| "loss": 3.2591, | |
| "step": 4300 | |
| }, | |
| { | |
| "epoch": 1.81, | |
| "eval_cer": 0.5522955691994837, | |
| "eval_loss": 1.7015308141708374, | |
| "eval_runtime": 353.5748, | |
| "eval_samples_per_second": 26.806, | |
| "eval_steps_per_second": 3.351, | |
| "step": 4300 | |
| }, | |
| { | |
| "epoch": 1.85, | |
| "grad_norm": 2.405174493789673, | |
| "learning_rate": 4.724876325088339e-05, | |
| "loss": 2.373, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 1.85, | |
| "eval_cer": 0.5220977865550819, | |
| "eval_loss": 2.095752477645874, | |
| "eval_runtime": 344.3236, | |
| "eval_samples_per_second": 27.526, | |
| "eval_steps_per_second": 3.442, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 1.89, | |
| "grad_norm": 3.823211908340454, | |
| "learning_rate": 4.717809187279153e-05, | |
| "loss": 2.3919, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 1.89, | |
| "eval_cer": 0.5331259043447656, | |
| "eval_loss": 2.41994571685791, | |
| "eval_runtime": 353.9416, | |
| "eval_samples_per_second": 26.778, | |
| "eval_steps_per_second": 3.348, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 1.94, | |
| "grad_norm": 2.0702412128448486, | |
| "learning_rate": 4.710742049469965e-05, | |
| "loss": 2.4842, | |
| "step": 4600 | |
| }, | |
| { | |
| "epoch": 1.94, | |
| "eval_cer": 0.5329523679167807, | |
| "eval_loss": 3.652872085571289, | |
| "eval_runtime": 339.8921, | |
| "eval_samples_per_second": 27.885, | |
| "eval_steps_per_second": 3.486, | |
| "step": 4600 | |
| }, | |
| { | |
| "epoch": 1.98, | |
| "grad_norm": 3.5580544471740723, | |
| "learning_rate": 4.703745583038869e-05, | |
| "loss": 3.577, | |
| "step": 4700 | |
| }, | |
| { | |
| "epoch": 1.98, | |
| "eval_cer": 0.5378284971256502, | |
| "eval_loss": 2.777487277984619, | |
| "eval_runtime": 356.1415, | |
| "eval_samples_per_second": 26.613, | |
| "eval_steps_per_second": 3.327, | |
| "step": 4700 | |
| }, | |
| { | |
| "epoch": 2.02, | |
| "grad_norm": 2.3930351734161377, | |
| "learning_rate": 4.6966784452296826e-05, | |
| "loss": 2.733, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 2.02, | |
| "eval_cer": 0.5305424113253294, | |
| "eval_loss": 3.2367630004882812, | |
| "eval_runtime": 346.0446, | |
| "eval_samples_per_second": 27.39, | |
| "eval_steps_per_second": 3.424, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 2.06, | |
| "grad_norm": 4.198298931121826, | |
| "learning_rate": 4.6896113074204954e-05, | |
| "loss": 1.616, | |
| "step": 4900 | |
| }, | |
| { | |
| "epoch": 2.06, | |
| "eval_cer": 0.523026573071057, | |
| "eval_loss": 2.150926351547241, | |
| "eval_runtime": 371.9418, | |
| "eval_samples_per_second": 25.482, | |
| "eval_steps_per_second": 3.186, | |
| "step": 4900 | |
| }, | |
| { | |
| "epoch": 2.1, | |
| "grad_norm": 3.6696269512176514, | |
| "learning_rate": 4.6825441696113075e-05, | |
| "loss": 1.8252, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 2.1, | |
| "eval_cer": 0.5257933792186461, | |
| "eval_loss": 3.1695752143859863, | |
| "eval_runtime": 351.2533, | |
| "eval_samples_per_second": 26.983, | |
| "eval_steps_per_second": 3.374, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 2.15, | |
| "grad_norm": 5.002595901489258, | |
| "learning_rate": 4.67547703180212e-05, | |
| "loss": 2.0594, | |
| "step": 5100 | |
| }, | |
| { | |
| "epoch": 2.15, | |
| "eval_cer": 0.5161682179030933, | |
| "eval_loss": 2.4667067527770996, | |
| "eval_runtime": 377.3711, | |
| "eval_samples_per_second": 25.116, | |
| "eval_steps_per_second": 3.14, | |
| "step": 5100 | |
| }, | |
| { | |
| "epoch": 2.19, | |
| "grad_norm": 3.2300162315368652, | |
| "learning_rate": 4.668409893992933e-05, | |
| "loss": 1.8512, | |
| "step": 5200 | |
| }, | |
| { | |
| "epoch": 2.19, | |
| "eval_cer": 0.5493185639982793, | |
| "eval_loss": 1.743632435798645, | |
| "eval_runtime": 361.1881, | |
| "eval_samples_per_second": 26.241, | |
| "eval_steps_per_second": 3.281, | |
| "step": 5200 | |
| }, | |
| { | |
| "epoch": 2.23, | |
| "grad_norm": 2.109961986541748, | |
| "learning_rate": 4.661342756183746e-05, | |
| "loss": 2.4585, | |
| "step": 5300 | |
| }, | |
| { | |
| "epoch": 2.23, | |
| "eval_cer": 0.5222273278323101, | |
| "eval_loss": 1.6703613996505737, | |
| "eval_runtime": 374.7053, | |
| "eval_samples_per_second": 25.295, | |
| "eval_steps_per_second": 3.162, | |
| "step": 5300 | |
| }, | |
| { | |
| "epoch": 2.27, | |
| "grad_norm": 3.7831201553344727, | |
| "learning_rate": 4.654275618374558e-05, | |
| "loss": 1.6107, | |
| "step": 5400 | |
| }, | |
| { | |
| "epoch": 2.27, | |
| "eval_cer": 0.5470601462594346, | |
| "eval_loss": 2.0870959758758545, | |
| "eval_runtime": 367.5169, | |
| "eval_samples_per_second": 25.789, | |
| "eval_steps_per_second": 3.224, | |
| "step": 5400 | |
| }, | |
| { | |
| "epoch": 2.32, | |
| "grad_norm": 12.008744239807129, | |
| "learning_rate": 4.6472084805653715e-05, | |
| "loss": 2.3883, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 2.32, | |
| "eval_cer": 0.5516771929138478, | |
| "eval_loss": 3.582747459411621, | |
| "eval_runtime": 376.938, | |
| "eval_samples_per_second": 25.145, | |
| "eval_steps_per_second": 3.144, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 2.36, | |
| "grad_norm": 2.4807026386260986, | |
| "learning_rate": 4.6401413427561844e-05, | |
| "loss": 2.1991, | |
| "step": 5600 | |
| }, | |
| { | |
| "epoch": 2.36, | |
| "eval_cer": 0.5373152203668218, | |
| "eval_loss": 3.5865066051483154, | |
| "eval_runtime": 360.1614, | |
| "eval_samples_per_second": 26.316, | |
| "eval_steps_per_second": 3.29, | |
| "step": 5600 | |
| }, | |
| { | |
| "epoch": 2.4, | |
| "grad_norm": 6.418155670166016, | |
| "learning_rate": 4.6330742049469965e-05, | |
| "loss": 2.3905, | |
| "step": 5700 | |
| }, | |
| { | |
| "epoch": 2.4, | |
| "eval_cer": 0.527748719252278, | |
| "eval_loss": 2.6406588554382324, | |
| "eval_runtime": 379.816, | |
| "eval_samples_per_second": 24.954, | |
| "eval_steps_per_second": 3.12, | |
| "step": 5700 | |
| }, | |
| { | |
| "epoch": 2.44, | |
| "grad_norm": 2.4806301593780518, | |
| "learning_rate": 4.626007067137809e-05, | |
| "loss": 3.3218, | |
| "step": 5800 | |
| }, | |
| { | |
| "epoch": 2.44, | |
| "eval_cer": 0.5217727112744907, | |
| "eval_loss": 1.8109639883041382, | |
| "eval_runtime": 351.8213, | |
| "eval_samples_per_second": 26.94, | |
| "eval_steps_per_second": 3.368, | |
| "step": 5800 | |
| }, | |
| { | |
| "epoch": 2.48, | |
| "grad_norm": 2.4776511192321777, | |
| "learning_rate": 4.618939929328622e-05, | |
| "loss": 2.3876, | |
| "step": 5900 | |
| }, | |
| { | |
| "epoch": 2.48, | |
| "eval_cer": 0.5089310156036134, | |
| "eval_loss": 3.134875774383545, | |
| "eval_runtime": 372.2436, | |
| "eval_samples_per_second": 25.462, | |
| "eval_steps_per_second": 3.183, | |
| "step": 5900 | |
| }, | |
| { | |
| "epoch": 2.53, | |
| "grad_norm": 9.97956657409668, | |
| "learning_rate": 4.611872791519435e-05, | |
| "loss": 4.1527, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 2.53, | |
| "eval_cer": 0.4981815337687224, | |
| "eval_loss": 1.7425427436828613, | |
| "eval_runtime": 363.658, | |
| "eval_samples_per_second": 26.063, | |
| "eval_steps_per_second": 3.259, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 2.57, | |
| "grad_norm": 3.353976249694824, | |
| "learning_rate": 4.604805653710248e-05, | |
| "loss": 1.7398, | |
| "step": 6100 | |
| }, | |
| { | |
| "epoch": 2.57, | |
| "eval_cer": 0.5255465175394001, | |
| "eval_loss": 2.6437551975250244, | |
| "eval_runtime": 380.4559, | |
| "eval_samples_per_second": 24.912, | |
| "eval_steps_per_second": 3.115, | |
| "step": 6100 | |
| }, | |
| { | |
| "epoch": 2.61, | |
| "grad_norm": 2.671917676925659, | |
| "learning_rate": 4.5978091872791526e-05, | |
| "loss": 2.2359, | |
| "step": 6200 | |
| }, | |
| { | |
| "epoch": 2.61, | |
| "eval_cer": 0.537452094169176, | |
| "eval_loss": 3.4509007930755615, | |
| "eval_runtime": 364.3511, | |
| "eval_samples_per_second": 26.013, | |
| "eval_steps_per_second": 3.252, | |
| "step": 6200 | |
| }, | |
| { | |
| "epoch": 2.65, | |
| "grad_norm": 10.607972145080566, | |
| "learning_rate": 4.590742049469965e-05, | |
| "loss": 3.6419, | |
| "step": 6300 | |
| }, | |
| { | |
| "epoch": 2.65, | |
| "eval_cer": 0.5374105431934614, | |
| "eval_loss": 2.2622523307800293, | |
| "eval_runtime": 370.8325, | |
| "eval_samples_per_second": 25.559, | |
| "eval_steps_per_second": 3.196, | |
| "step": 6300 | |
| }, | |
| { | |
| "epoch": 2.69, | |
| "grad_norm": 4.678300857543945, | |
| "learning_rate": 4.5836749116607775e-05, | |
| "loss": 1.7309, | |
| "step": 6400 | |
| }, | |
| { | |
| "epoch": 2.69, | |
| "eval_cer": 0.5418393883696375, | |
| "eval_loss": 1.8007577657699585, | |
| "eval_runtime": 357.8741, | |
| "eval_samples_per_second": 26.484, | |
| "eval_steps_per_second": 3.311, | |
| "step": 6400 | |
| }, | |
| { | |
| "epoch": 2.74, | |
| "grad_norm": 3.6259689331054688, | |
| "learning_rate": 4.57660777385159e-05, | |
| "loss": 2.784, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 2.74, | |
| "eval_cer": 0.5228799225685347, | |
| "eval_loss": 2.9312586784362793, | |
| "eval_runtime": 371.6433, | |
| "eval_samples_per_second": 25.503, | |
| "eval_steps_per_second": 3.189, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 2.78, | |
| "grad_norm": 1.8296568393707275, | |
| "learning_rate": 4.569540636042403e-05, | |
| "loss": 2.4269, | |
| "step": 6600 | |
| }, | |
| { | |
| "epoch": 2.78, | |
| "eval_cer": 0.5211445582886864, | |
| "eval_loss": 3.522822618484497, | |
| "eval_runtime": 359.0335, | |
| "eval_samples_per_second": 26.399, | |
| "eval_steps_per_second": 3.301, | |
| "step": 6600 | |
| }, | |
| { | |
| "epoch": 2.82, | |
| "grad_norm": 23.777395248413086, | |
| "learning_rate": 4.562473498233216e-05, | |
| "loss": 1.5294, | |
| "step": 6700 | |
| }, | |
| { | |
| "epoch": 2.82, | |
| "eval_cer": 0.5097864768683275, | |
| "eval_loss": 2.758403778076172, | |
| "eval_runtime": 375.8984, | |
| "eval_samples_per_second": 25.214, | |
| "eval_steps_per_second": 3.152, | |
| "step": 6700 | |
| }, | |
| { | |
| "epoch": 2.86, | |
| "grad_norm": 5.520616054534912, | |
| "learning_rate": 4.555406360424028e-05, | |
| "loss": 1.6177, | |
| "step": 6800 | |
| }, | |
| { | |
| "epoch": 2.86, | |
| "eval_cer": 0.5417905048687967, | |
| "eval_loss": 2.8678946495056152, | |
| "eval_runtime": 360.269, | |
| "eval_samples_per_second": 26.308, | |
| "eval_steps_per_second": 3.289, | |
| "step": 6800 | |
| }, | |
| { | |
| "epoch": 2.9, | |
| "grad_norm": 12.704540252685547, | |
| "learning_rate": 4.5483392226148416e-05, | |
| "loss": 2.6892, | |
| "step": 6900 | |
| }, | |
| { | |
| "epoch": 2.9, | |
| "eval_cer": 0.5306426225020531, | |
| "eval_loss": 2.7854502201080322, | |
| "eval_runtime": 372.5558, | |
| "eval_samples_per_second": 25.44, | |
| "eval_steps_per_second": 3.181, | |
| "step": 6900 | |
| }, | |
| { | |
| "epoch": 2.95, | |
| "grad_norm": 3.1302738189697266, | |
| "learning_rate": 4.541272084805654e-05, | |
| "loss": 3.1467, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 2.95, | |
| "eval_cer": 0.5148948027061906, | |
| "eval_loss": 2.9890220165252686, | |
| "eval_runtime": 359.2287, | |
| "eval_samples_per_second": 26.384, | |
| "eval_steps_per_second": 3.299, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 2.99, | |
| "grad_norm": 1.750848412513733, | |
| "learning_rate": 4.5342049469964665e-05, | |
| "loss": 2.1825, | |
| "step": 7100 | |
| }, | |
| { | |
| "epoch": 2.99, | |
| "eval_cer": 0.5565606546478432, | |
| "eval_loss": 3.195390224456787, | |
| "eval_runtime": 366.6828, | |
| "eval_samples_per_second": 25.848, | |
| "eval_steps_per_second": 3.232, | |
| "step": 7100 | |
| } | |
| ], | |
| "logging_steps": 100, | |
| "max_steps": 71250, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 30, | |
| "save_steps": 100, | |
| "total_flos": 7.774285790233235e+19, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |