| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 49.079754601226995, | |
| "global_step": 16000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.23, | |
| "learning_rate": 0.0002376, | |
| "loss": 6.548, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.23, | |
| "eval_cer": 1.0, | |
| "eval_loss": 3.476283073425293, | |
| "eval_runtime": 60.3486, | |
| "eval_samples_per_second": 26.513, | |
| "eval_steps_per_second": 1.657, | |
| "eval_wer": 1.0, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 2.45, | |
| "learning_rate": 0.000294379746835443, | |
| "loss": 3.42, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.45, | |
| "eval_cer": 1.0, | |
| "eval_loss": 3.3155558109283447, | |
| "eval_runtime": 60.3859, | |
| "eval_samples_per_second": 26.496, | |
| "eval_steps_per_second": 1.656, | |
| "eval_wer": 1.0, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 3.68, | |
| "learning_rate": 0.00028678481012658224, | |
| "loss": 3.291, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 3.68, | |
| "eval_cer": 1.0, | |
| "eval_loss": 3.2396233081817627, | |
| "eval_runtime": 60.8505, | |
| "eval_samples_per_second": 26.294, | |
| "eval_steps_per_second": 1.643, | |
| "eval_wer": 1.0, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 4.91, | |
| "learning_rate": 0.0002791898734177215, | |
| "loss": 2.6515, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 4.91, | |
| "eval_cer": 0.5835108495504928, | |
| "eval_loss": 2.042246103286743, | |
| "eval_runtime": 60.4335, | |
| "eval_samples_per_second": 26.475, | |
| "eval_steps_per_second": 1.655, | |
| "eval_wer": 0.999671772428884, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 6.13, | |
| "learning_rate": 0.00027159493670886076, | |
| "loss": 1.7019, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 6.13, | |
| "eval_cer": 0.4797450987471567, | |
| "eval_loss": 1.6337121725082397, | |
| "eval_runtime": 60.2878, | |
| "eval_samples_per_second": 26.539, | |
| "eval_steps_per_second": 1.659, | |
| "eval_wer": 0.9892778993435448, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 7.36, | |
| "learning_rate": 0.00026399999999999997, | |
| "loss": 1.3604, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 7.36, | |
| "eval_cer": 0.44627576993898255, | |
| "eval_loss": 1.5220645666122437, | |
| "eval_runtime": 60.8256, | |
| "eval_samples_per_second": 26.305, | |
| "eval_steps_per_second": 1.644, | |
| "eval_wer": 0.987527352297593, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 8.59, | |
| "learning_rate": 0.0002564050632911392, | |
| "loss": 1.1965, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 8.59, | |
| "eval_cer": 0.4246669314366177, | |
| "eval_loss": 1.528436303138733, | |
| "eval_runtime": 60.5379, | |
| "eval_samples_per_second": 26.43, | |
| "eval_steps_per_second": 1.652, | |
| "eval_wer": 0.9765864332603938, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 9.82, | |
| "learning_rate": 0.00024881012658227843, | |
| "loss": 1.069, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 9.82, | |
| "eval_cer": 0.41239123370762176, | |
| "eval_loss": 1.5227978229522705, | |
| "eval_runtime": 60.2014, | |
| "eval_samples_per_second": 26.577, | |
| "eval_steps_per_second": 1.661, | |
| "eval_wer": 0.9671772428884027, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 11.04, | |
| "learning_rate": 0.00024121518987341772, | |
| "loss": 0.9536, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 11.04, | |
| "eval_cer": 0.38684695093331406, | |
| "eval_loss": 1.4059038162231445, | |
| "eval_runtime": 60.2947, | |
| "eval_samples_per_second": 26.536, | |
| "eval_steps_per_second": 1.659, | |
| "eval_wer": 0.9599562363238512, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 12.27, | |
| "learning_rate": 0.00023362025316455695, | |
| "loss": 0.8487, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 12.27, | |
| "eval_cer": 0.37390331082788747, | |
| "eval_loss": 1.4082870483398438, | |
| "eval_runtime": 60.1938, | |
| "eval_samples_per_second": 26.581, | |
| "eval_steps_per_second": 1.661, | |
| "eval_wer": 0.950109409190372, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 13.5, | |
| "learning_rate": 0.00022602531645569618, | |
| "loss": 0.7655, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 13.5, | |
| "eval_cer": 0.36121240567570495, | |
| "eval_loss": 1.40787935256958, | |
| "eval_runtime": 60.2423, | |
| "eval_samples_per_second": 26.559, | |
| "eval_steps_per_second": 1.66, | |
| "eval_wer": 0.9368708971553611, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 14.72, | |
| "learning_rate": 0.0002184303797468354, | |
| "loss": 0.6956, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 14.72, | |
| "eval_cer": 0.34590388850778064, | |
| "eval_loss": 1.417035698890686, | |
| "eval_runtime": 60.1538, | |
| "eval_samples_per_second": 26.598, | |
| "eval_steps_per_second": 1.662, | |
| "eval_wer": 0.9411378555798687, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 15.95, | |
| "learning_rate": 0.00021083544303797464, | |
| "loss": 0.6287, | |
| "step": 5200 | |
| }, | |
| { | |
| "epoch": 15.95, | |
| "eval_cer": 0.3383579449037802, | |
| "eval_loss": 1.3999766111373901, | |
| "eval_runtime": 60.2038, | |
| "eval_samples_per_second": 26.576, | |
| "eval_steps_per_second": 1.661, | |
| "eval_wer": 0.9235229759299781, | |
| "step": 5200 | |
| }, | |
| { | |
| "epoch": 17.18, | |
| "learning_rate": 0.00020324050632911393, | |
| "loss": 0.561, | |
| "step": 5600 | |
| }, | |
| { | |
| "epoch": 17.18, | |
| "eval_cer": 0.32953027403689933, | |
| "eval_loss": 1.4734982252120972, | |
| "eval_runtime": 60.3387, | |
| "eval_samples_per_second": 26.517, | |
| "eval_steps_per_second": 1.657, | |
| "eval_wer": 0.9022975929978119, | |
| "step": 5600 | |
| }, | |
| { | |
| "epoch": 18.4, | |
| "learning_rate": 0.00019564556962025316, | |
| "loss": 0.5155, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 18.4, | |
| "eval_cer": 0.32232732786944435, | |
| "eval_loss": 1.538634181022644, | |
| "eval_runtime": 69.1066, | |
| "eval_samples_per_second": 23.153, | |
| "eval_steps_per_second": 1.447, | |
| "eval_wer": 0.9202407002188184, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 19.63, | |
| "learning_rate": 0.0001880506329113924, | |
| "loss": 0.4864, | |
| "step": 6400 | |
| }, | |
| { | |
| "epoch": 19.63, | |
| "eval_cer": 0.32590172220818137, | |
| "eval_loss": 1.618619680404663, | |
| "eval_runtime": 69.0557, | |
| "eval_samples_per_second": 23.17, | |
| "eval_steps_per_second": 1.448, | |
| "eval_wer": 0.9073304157549235, | |
| "step": 6400 | |
| }, | |
| { | |
| "epoch": 20.86, | |
| "learning_rate": 0.00018045569620253163, | |
| "loss": 0.4261, | |
| "step": 6800 | |
| }, | |
| { | |
| "epoch": 20.86, | |
| "eval_cer": 0.313030292089396, | |
| "eval_loss": 1.6417021751403809, | |
| "eval_runtime": 68.875, | |
| "eval_samples_per_second": 23.23, | |
| "eval_steps_per_second": 1.452, | |
| "eval_wer": 0.9216630196936543, | |
| "step": 6800 | |
| }, | |
| { | |
| "epoch": 22.09, | |
| "learning_rate": 0.00017286075949367088, | |
| "loss": 0.4051, | |
| "step": 7200 | |
| }, | |
| { | |
| "epoch": 22.09, | |
| "eval_cer": 0.30263205401307003, | |
| "eval_loss": 1.6295086145401, | |
| "eval_runtime": 69.3362, | |
| "eval_samples_per_second": 23.076, | |
| "eval_steps_per_second": 1.442, | |
| "eval_wer": 0.8954048140043763, | |
| "step": 7200 | |
| }, | |
| { | |
| "epoch": 23.31, | |
| "learning_rate": 0.00016526582278481012, | |
| "loss": 0.3779, | |
| "step": 7600 | |
| }, | |
| { | |
| "epoch": 23.31, | |
| "eval_cer": 0.31534101166191286, | |
| "eval_loss": 1.8218317031860352, | |
| "eval_runtime": 68.8091, | |
| "eval_samples_per_second": 23.253, | |
| "eval_steps_per_second": 1.453, | |
| "eval_wer": 0.8979212253829322, | |
| "step": 7600 | |
| }, | |
| { | |
| "epoch": 24.54, | |
| "learning_rate": 0.00015767088607594935, | |
| "loss": 0.35, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 24.54, | |
| "eval_cer": 0.303570783839405, | |
| "eval_loss": 1.779032588005066, | |
| "eval_runtime": 68.3457, | |
| "eval_samples_per_second": 23.41, | |
| "eval_steps_per_second": 1.463, | |
| "eval_wer": 0.8921225382932166, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 25.77, | |
| "learning_rate": 0.00015007594936708858, | |
| "loss": 0.3343, | |
| "step": 8400 | |
| }, | |
| { | |
| "epoch": 25.77, | |
| "eval_cer": 0.3071812831714626, | |
| "eval_loss": 1.8588141202926636, | |
| "eval_runtime": 60.0463, | |
| "eval_samples_per_second": 26.646, | |
| "eval_steps_per_second": 1.665, | |
| "eval_wer": 0.9113785557986871, | |
| "step": 8400 | |
| }, | |
| { | |
| "epoch": 26.99, | |
| "learning_rate": 0.00014248101265822784, | |
| "loss": 0.3137, | |
| "step": 8800 | |
| }, | |
| { | |
| "epoch": 26.99, | |
| "eval_cer": 0.29351554319962453, | |
| "eval_loss": 1.8095606565475464, | |
| "eval_runtime": 59.7272, | |
| "eval_samples_per_second": 26.788, | |
| "eval_steps_per_second": 1.674, | |
| "eval_wer": 0.8756017505470459, | |
| "step": 8800 | |
| }, | |
| { | |
| "epoch": 28.22, | |
| "learning_rate": 0.00013488607594936707, | |
| "loss": 0.299, | |
| "step": 9200 | |
| }, | |
| { | |
| "epoch": 28.22, | |
| "eval_cer": 0.3023251615698451, | |
| "eval_loss": 1.9720637798309326, | |
| "eval_runtime": 60.067, | |
| "eval_samples_per_second": 26.637, | |
| "eval_steps_per_second": 1.665, | |
| "eval_wer": 0.8863238512035011, | |
| "step": 9200 | |
| }, | |
| { | |
| "epoch": 29.45, | |
| "learning_rate": 0.00012729113924050633, | |
| "loss": 0.2894, | |
| "step": 9600 | |
| }, | |
| { | |
| "epoch": 29.45, | |
| "eval_cer": 0.29584431526880167, | |
| "eval_loss": 1.9907439947128296, | |
| "eval_runtime": 68.8361, | |
| "eval_samples_per_second": 23.244, | |
| "eval_steps_per_second": 1.453, | |
| "eval_wer": 0.887199124726477, | |
| "step": 9600 | |
| }, | |
| { | |
| "epoch": 30.67, | |
| "learning_rate": 0.00011969620253164556, | |
| "loss": 0.2784, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 30.67, | |
| "eval_cer": 0.2944542730259595, | |
| "eval_loss": 1.9494301080703735, | |
| "eval_runtime": 68.9865, | |
| "eval_samples_per_second": 23.193, | |
| "eval_steps_per_second": 1.45, | |
| "eval_wer": 0.9089715536105033, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 31.9, | |
| "learning_rate": 0.00011210126582278481, | |
| "loss": 0.2662, | |
| "step": 10400 | |
| }, | |
| { | |
| "epoch": 31.9, | |
| "eval_cer": 0.29346138570964364, | |
| "eval_loss": 1.995172142982483, | |
| "eval_runtime": 68.9796, | |
| "eval_samples_per_second": 23.195, | |
| "eval_steps_per_second": 1.45, | |
| "eval_wer": 0.8978118161925602, | |
| "step": 10400 | |
| }, | |
| { | |
| "epoch": 33.13, | |
| "learning_rate": 0.00010450632911392404, | |
| "loss": 0.2614, | |
| "step": 10800 | |
| }, | |
| { | |
| "epoch": 33.13, | |
| "eval_cer": 0.29790229988807454, | |
| "eval_loss": 2.0600392818450928, | |
| "eval_runtime": 68.6491, | |
| "eval_samples_per_second": 23.307, | |
| "eval_steps_per_second": 1.457, | |
| "eval_wer": 0.8948577680525164, | |
| "step": 10800 | |
| }, | |
| { | |
| "epoch": 34.36, | |
| "learning_rate": 9.691139240506327e-05, | |
| "loss": 0.2401, | |
| "step": 11200 | |
| }, | |
| { | |
| "epoch": 34.36, | |
| "eval_cer": 0.29495974293244753, | |
| "eval_loss": 2.118035316467285, | |
| "eval_runtime": 68.8384, | |
| "eval_samples_per_second": 23.243, | |
| "eval_steps_per_second": 1.453, | |
| "eval_wer": 0.8913566739606127, | |
| "step": 11200 | |
| }, | |
| { | |
| "epoch": 35.58, | |
| "learning_rate": 8.933544303797467e-05, | |
| "loss": 0.2392, | |
| "step": 11600 | |
| }, | |
| { | |
| "epoch": 35.58, | |
| "eval_cer": 0.28950788894104057, | |
| "eval_loss": 2.1196768283843994, | |
| "eval_runtime": 68.573, | |
| "eval_samples_per_second": 23.333, | |
| "eval_steps_per_second": 1.458, | |
| "eval_wer": 0.8713347921225383, | |
| "step": 11600 | |
| }, | |
| { | |
| "epoch": 36.81, | |
| "learning_rate": 8.175949367088606e-05, | |
| "loss": 0.23, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 36.81, | |
| "eval_cer": 0.2940932230927537, | |
| "eval_loss": 2.168043851852417, | |
| "eval_runtime": 68.6136, | |
| "eval_samples_per_second": 23.319, | |
| "eval_steps_per_second": 1.457, | |
| "eval_wer": 0.8713347921225383, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 38.04, | |
| "learning_rate": 7.416455696202532e-05, | |
| "loss": 0.2246, | |
| "step": 12400 | |
| }, | |
| { | |
| "epoch": 38.04, | |
| "eval_cer": 0.2879192692349352, | |
| "eval_loss": 2.1525843143463135, | |
| "eval_runtime": 59.7644, | |
| "eval_samples_per_second": 26.772, | |
| "eval_steps_per_second": 1.673, | |
| "eval_wer": 0.874070021881838, | |
| "step": 12400 | |
| }, | |
| { | |
| "epoch": 39.26, | |
| "learning_rate": 6.656962025316455e-05, | |
| "loss": 0.2152, | |
| "step": 12800 | |
| }, | |
| { | |
| "epoch": 39.26, | |
| "eval_cer": 0.28893020904791133, | |
| "eval_loss": 2.263143301010132, | |
| "eval_runtime": 60.4724, | |
| "eval_samples_per_second": 26.458, | |
| "eval_steps_per_second": 1.654, | |
| "eval_wer": 0.8789934354485777, | |
| "step": 12800 | |
| }, | |
| { | |
| "epoch": 40.49, | |
| "learning_rate": 5.897468354430379e-05, | |
| "loss": 0.212, | |
| "step": 13200 | |
| }, | |
| { | |
| "epoch": 40.49, | |
| "eval_cer": 0.28430876990287757, | |
| "eval_loss": 2.2723913192749023, | |
| "eval_runtime": 60.2908, | |
| "eval_samples_per_second": 26.538, | |
| "eval_steps_per_second": 1.659, | |
| "eval_wer": 0.8660831509846827, | |
| "step": 13200 | |
| }, | |
| { | |
| "epoch": 41.72, | |
| "learning_rate": 5.1379746835443034e-05, | |
| "loss": 0.2044, | |
| "step": 13600 | |
| }, | |
| { | |
| "epoch": 41.72, | |
| "eval_cer": 0.2877929017583132, | |
| "eval_loss": 2.2438297271728516, | |
| "eval_runtime": 70.0914, | |
| "eval_samples_per_second": 22.827, | |
| "eval_steps_per_second": 1.427, | |
| "eval_wer": 0.8691466083150985, | |
| "step": 13600 | |
| }, | |
| { | |
| "epoch": 42.94, | |
| "learning_rate": 4.380379746835442e-05, | |
| "loss": 0.2029, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 42.94, | |
| "eval_cer": 0.28327977759324113, | |
| "eval_loss": 2.2518999576568604, | |
| "eval_runtime": 69.2505, | |
| "eval_samples_per_second": 23.105, | |
| "eval_steps_per_second": 1.444, | |
| "eval_wer": 0.8576586433260394, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 44.17, | |
| "learning_rate": 3.6208860759493666e-05, | |
| "loss": 0.1972, | |
| "step": 14400 | |
| }, | |
| { | |
| "epoch": 44.17, | |
| "eval_cer": 0.28133010795393004, | |
| "eval_loss": 2.2697150707244873, | |
| "eval_runtime": 68.5533, | |
| "eval_samples_per_second": 23.34, | |
| "eval_steps_per_second": 1.459, | |
| "eval_wer": 0.8603938730853392, | |
| "step": 14400 | |
| }, | |
| { | |
| "epoch": 45.4, | |
| "learning_rate": 2.861392405063291e-05, | |
| "loss": 0.1884, | |
| "step": 14800 | |
| }, | |
| { | |
| "epoch": 45.4, | |
| "eval_cer": 0.2846878723327436, | |
| "eval_loss": 2.3294308185577393, | |
| "eval_runtime": 60.5028, | |
| "eval_samples_per_second": 26.445, | |
| "eval_steps_per_second": 1.653, | |
| "eval_wer": 0.8661925601750547, | |
| "step": 14800 | |
| }, | |
| { | |
| "epoch": 46.63, | |
| "learning_rate": 2.101898734177215e-05, | |
| "loss": 0.1877, | |
| "step": 15200 | |
| }, | |
| { | |
| "epoch": 46.63, | |
| "eval_cer": 0.2792721233346572, | |
| "eval_loss": 2.3077094554901123, | |
| "eval_runtime": 60.8923, | |
| "eval_samples_per_second": 26.276, | |
| "eval_steps_per_second": 1.642, | |
| "eval_wer": 0.8561269146608315, | |
| "step": 15200 | |
| }, | |
| { | |
| "epoch": 47.85, | |
| "learning_rate": 1.3424050632911391e-05, | |
| "loss": 0.1871, | |
| "step": 15600 | |
| }, | |
| { | |
| "epoch": 47.85, | |
| "eval_cer": 0.2801025381810304, | |
| "eval_loss": 2.351794481277466, | |
| "eval_runtime": 60.3104, | |
| "eval_samples_per_second": 26.529, | |
| "eval_steps_per_second": 1.658, | |
| "eval_wer": 0.8563457330415755, | |
| "step": 15600 | |
| }, | |
| { | |
| "epoch": 49.08, | |
| "learning_rate": 5.829113924050632e-06, | |
| "loss": 0.1838, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 49.08, | |
| "eval_cer": 0.27992201321442756, | |
| "eval_loss": 2.3462095260620117, | |
| "eval_runtime": 60.2933, | |
| "eval_samples_per_second": 26.537, | |
| "eval_steps_per_second": 1.659, | |
| "eval_wer": 0.8555798687089715, | |
| "step": 16000 | |
| } | |
| ], | |
| "max_steps": 16300, | |
| "num_train_epochs": 50, | |
| "total_flos": 6.654066355887293e+19, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |