| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 23.9993690851735, | |
| "global_step": 19008, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.5, | |
| "learning_rate": 0.00023999999999999998, | |
| "loss": 46.1613, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "eval_cer": 1.0, | |
| "eval_loss": 6.5866594314575195, | |
| "eval_runtime": 34.0665, | |
| "eval_samples_per_second": 15.176, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.01, | |
| "learning_rate": 0.00029769820971867005, | |
| "loss": 6.421, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 1.01, | |
| "eval_cer": 1.0, | |
| "eval_loss": 6.284526348114014, | |
| "eval_runtime": 23.4938, | |
| "eval_samples_per_second": 22.006, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 1.51, | |
| "learning_rate": 0.00029462915601023015, | |
| "loss": 6.2208, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 1.51, | |
| "eval_cer": 1.0, | |
| "eval_loss": 6.20452356338501, | |
| "eval_runtime": 21.4276, | |
| "eval_samples_per_second": 24.128, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 2.02, | |
| "learning_rate": 0.00029156010230179026, | |
| "loss": 5.8856, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 2.02, | |
| "eval_cer": 0.8329783457578984, | |
| "eval_loss": 4.5766448974609375, | |
| "eval_runtime": 22.0378, | |
| "eval_samples_per_second": 23.46, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 2.52, | |
| "learning_rate": 0.00028849104859335036, | |
| "loss": 3.6701, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.52, | |
| "eval_cer": 0.6766063187788427, | |
| "eval_loss": 2.7600414752960205, | |
| "eval_runtime": 32.6882, | |
| "eval_samples_per_second": 15.816, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 3.03, | |
| "learning_rate": 0.00028542199488491046, | |
| "loss": 2.5771, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 3.03, | |
| "eval_cer": 0.580937167199148, | |
| "eval_loss": 2.0640788078308105, | |
| "eval_runtime": 22.8394, | |
| "eval_samples_per_second": 22.636, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 3.54, | |
| "learning_rate": 0.00028235294117647056, | |
| "loss": 2.0353, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 3.54, | |
| "eval_cer": 0.4971600993965211, | |
| "eval_loss": 1.7469913959503174, | |
| "eval_runtime": 22.1406, | |
| "eval_samples_per_second": 23.351, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 4.04, | |
| "learning_rate": 0.00027928388746803067, | |
| "loss": 1.7298, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 4.04, | |
| "eval_cer": 0.4518991835285765, | |
| "eval_loss": 1.5479540824890137, | |
| "eval_runtime": 22.65, | |
| "eval_samples_per_second": 22.826, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 4.55, | |
| "learning_rate": 0.00027621483375959077, | |
| "loss": 1.4816, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 4.55, | |
| "eval_cer": 0.4222577209797657, | |
| "eval_loss": 1.4363528490066528, | |
| "eval_runtime": 22.4962, | |
| "eval_samples_per_second": 22.982, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 5.05, | |
| "learning_rate": 0.00027314578005115087, | |
| "loss": 1.3714, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 5.05, | |
| "eval_cer": 0.3995385161519347, | |
| "eval_loss": 1.302014946937561, | |
| "eval_runtime": 22.4129, | |
| "eval_samples_per_second": 23.067, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 5.56, | |
| "learning_rate": 0.000270076726342711, | |
| "loss": 1.2056, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 5.56, | |
| "eval_cer": 0.39527866524671634, | |
| "eval_loss": 1.2633957862854004, | |
| "eval_runtime": 22.284, | |
| "eval_samples_per_second": 23.201, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 6.06, | |
| "learning_rate": 0.0002670076726342711, | |
| "loss": 1.1594, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 6.06, | |
| "eval_cer": 0.39261625843095493, | |
| "eval_loss": 1.2651187181472778, | |
| "eval_runtime": 22.9096, | |
| "eval_samples_per_second": 22.567, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 6.57, | |
| "learning_rate": 0.0002639386189258312, | |
| "loss": 1.0238, | |
| "step": 5200 | |
| }, | |
| { | |
| "epoch": 6.57, | |
| "eval_cer": 0.37149449769258075, | |
| "eval_loss": 1.1931949853897095, | |
| "eval_runtime": 21.8388, | |
| "eval_samples_per_second": 23.673, | |
| "step": 5200 | |
| }, | |
| { | |
| "epoch": 7.07, | |
| "learning_rate": 0.0002608695652173913, | |
| "loss": 1.0155, | |
| "step": 5600 | |
| }, | |
| { | |
| "epoch": 7.07, | |
| "eval_cer": 0.365814696485623, | |
| "eval_loss": 1.1498987674713135, | |
| "eval_runtime": 22.3033, | |
| "eval_samples_per_second": 23.18, | |
| "step": 5600 | |
| }, | |
| { | |
| "epoch": 7.58, | |
| "learning_rate": 0.0002578005115089514, | |
| "loss": 0.9235, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 7.58, | |
| "eval_cer": 0.35197018104366345, | |
| "eval_loss": 1.1570812463760376, | |
| "eval_runtime": 21.7227, | |
| "eval_samples_per_second": 23.8, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 8.08, | |
| "learning_rate": 0.0002547314578005115, | |
| "loss": 0.8688, | |
| "step": 6400 | |
| }, | |
| { | |
| "epoch": 8.08, | |
| "eval_cer": 0.3510827121050763, | |
| "eval_loss": 1.1230961084365845, | |
| "eval_runtime": 21.7451, | |
| "eval_samples_per_second": 23.775, | |
| "step": 6400 | |
| }, | |
| { | |
| "epoch": 8.59, | |
| "learning_rate": 0.0002516624040920716, | |
| "loss": 0.7623, | |
| "step": 6800 | |
| }, | |
| { | |
| "epoch": 8.59, | |
| "eval_cer": 0.3438054668086617, | |
| "eval_loss": 1.096892237663269, | |
| "eval_runtime": 21.8281, | |
| "eval_samples_per_second": 23.685, | |
| "step": 6800 | |
| }, | |
| { | |
| "epoch": 9.09, | |
| "learning_rate": 0.0002485933503836317, | |
| "loss": 0.7928, | |
| "step": 7200 | |
| }, | |
| { | |
| "epoch": 9.09, | |
| "eval_cer": 0.35640752573659923, | |
| "eval_loss": 1.176791787147522, | |
| "eval_runtime": 23.1776, | |
| "eval_samples_per_second": 22.306, | |
| "step": 7200 | |
| }, | |
| { | |
| "epoch": 9.6, | |
| "learning_rate": 0.0002455242966751918, | |
| "loss": 0.7114, | |
| "step": 7600 | |
| }, | |
| { | |
| "epoch": 9.6, | |
| "eval_cer": 0.3354632587859425, | |
| "eval_loss": 1.0966300964355469, | |
| "eval_runtime": 22.9565, | |
| "eval_samples_per_second": 22.521, | |
| "step": 7600 | |
| }, | |
| { | |
| "epoch": 10.1, | |
| "learning_rate": 0.0002424552429667519, | |
| "loss": 0.6637, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 10.1, | |
| "eval_cer": 0.3336883209087682, | |
| "eval_loss": 1.115621566772461, | |
| "eval_runtime": 22.4374, | |
| "eval_samples_per_second": 23.042, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 10.61, | |
| "learning_rate": 0.000239386189258312, | |
| "loss": 0.594, | |
| "step": 8400 | |
| }, | |
| { | |
| "epoch": 10.61, | |
| "eval_cer": 0.33457578984735536, | |
| "eval_loss": 1.1540485620498657, | |
| "eval_runtime": 23.8417, | |
| "eval_samples_per_second": 21.685, | |
| "step": 8400 | |
| }, | |
| { | |
| "epoch": 11.11, | |
| "learning_rate": 0.0002363171355498721, | |
| "loss": 0.6477, | |
| "step": 8800 | |
| }, | |
| { | |
| "epoch": 11.11, | |
| "eval_cer": 0.34593539226127085, | |
| "eval_loss": 1.1207655668258667, | |
| "eval_runtime": 24.0744, | |
| "eval_samples_per_second": 21.475, | |
| "step": 8800 | |
| }, | |
| { | |
| "epoch": 11.62, | |
| "learning_rate": 0.00023324808184143218, | |
| "loss": 0.5561, | |
| "step": 9200 | |
| }, | |
| { | |
| "epoch": 11.62, | |
| "eval_cer": 0.327831025914093, | |
| "eval_loss": 1.070268988609314, | |
| "eval_runtime": 30.8402, | |
| "eval_samples_per_second": 16.764, | |
| "step": 9200 | |
| }, | |
| { | |
| "epoch": 12.12, | |
| "learning_rate": 0.0002301790281329923, | |
| "loss": 0.5367, | |
| "step": 9600 | |
| }, | |
| { | |
| "epoch": 12.12, | |
| "eval_cer": 0.32197373091941783, | |
| "eval_loss": 1.0586289167404175, | |
| "eval_runtime": 23.6938, | |
| "eval_samples_per_second": 21.82, | |
| "step": 9600 | |
| }, | |
| { | |
| "epoch": 12.63, | |
| "learning_rate": 0.0002271099744245524, | |
| "loss": 0.4913, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 12.63, | |
| "eval_cer": 0.3210862619808307, | |
| "eval_loss": 1.050113558769226, | |
| "eval_runtime": 23.0884, | |
| "eval_samples_per_second": 22.392, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 13.13, | |
| "learning_rate": 0.0002240409207161125, | |
| "loss": 0.477, | |
| "step": 10400 | |
| }, | |
| { | |
| "epoch": 13.13, | |
| "eval_cer": 0.3170039048633298, | |
| "eval_loss": 1.076606273651123, | |
| "eval_runtime": 22.7762, | |
| "eval_samples_per_second": 22.699, | |
| "step": 10400 | |
| }, | |
| { | |
| "epoch": 13.64, | |
| "learning_rate": 0.00022097186700767261, | |
| "loss": 0.43, | |
| "step": 10800 | |
| }, | |
| { | |
| "epoch": 13.64, | |
| "eval_cer": 0.31576144834930775, | |
| "eval_loss": 1.0695208311080933, | |
| "eval_runtime": 23.2744, | |
| "eval_samples_per_second": 22.213, | |
| "step": 10800 | |
| }, | |
| { | |
| "epoch": 14.14, | |
| "learning_rate": 0.0002179028132992327, | |
| "loss": 0.4295, | |
| "step": 11200 | |
| }, | |
| { | |
| "epoch": 14.14, | |
| "eval_cer": 0.30670926517571884, | |
| "eval_loss": 1.0685380697250366, | |
| "eval_runtime": 22.189, | |
| "eval_samples_per_second": 23.3, | |
| "step": 11200 | |
| }, | |
| { | |
| "epoch": 14.65, | |
| "learning_rate": 0.00021483375959079282, | |
| "loss": 0.3727, | |
| "step": 11600 | |
| }, | |
| { | |
| "epoch": 14.65, | |
| "eval_cer": 0.3118565850195243, | |
| "eval_loss": 1.070268988609314, | |
| "eval_runtime": 23.0689, | |
| "eval_samples_per_second": 22.411, | |
| "step": 11600 | |
| }, | |
| { | |
| "epoch": 15.15, | |
| "learning_rate": 0.00021176470588235295, | |
| "loss": 0.3726, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 15.15, | |
| "eval_cer": 0.3095491657791977, | |
| "eval_loss": 1.0649579763412476, | |
| "eval_runtime": 24.1518, | |
| "eval_samples_per_second": 21.406, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 15.66, | |
| "learning_rate": 0.00020869565217391303, | |
| "loss": 0.3471, | |
| "step": 12400 | |
| }, | |
| { | |
| "epoch": 15.66, | |
| "eval_cer": 0.3139865104721335, | |
| "eval_loss": 1.0912107229232788, | |
| "eval_runtime": 22.5301, | |
| "eval_samples_per_second": 22.947, | |
| "step": 12400 | |
| }, | |
| { | |
| "epoch": 16.16, | |
| "learning_rate": 0.00020562659846547313, | |
| "loss": 0.3519, | |
| "step": 12800 | |
| }, | |
| { | |
| "epoch": 16.16, | |
| "eval_cer": 0.31274405395811145, | |
| "eval_loss": 1.110485315322876, | |
| "eval_runtime": 21.9816, | |
| "eval_samples_per_second": 23.52, | |
| "step": 12800 | |
| }, | |
| { | |
| "epoch": 16.67, | |
| "learning_rate": 0.00020255754475703323, | |
| "loss": 0.3356, | |
| "step": 13200 | |
| }, | |
| { | |
| "epoch": 16.67, | |
| "eval_cer": 0.307596734114306, | |
| "eval_loss": 1.111737608909607, | |
| "eval_runtime": 21.9913, | |
| "eval_samples_per_second": 23.509, | |
| "step": 13200 | |
| }, | |
| { | |
| "epoch": 17.17, | |
| "learning_rate": 0.00019948849104859333, | |
| "loss": 0.3116, | |
| "step": 13600 | |
| }, | |
| { | |
| "epoch": 17.17, | |
| "eval_cer": 0.3049343272985445, | |
| "eval_loss": 1.0982708930969238, | |
| "eval_runtime": 21.5115, | |
| "eval_samples_per_second": 24.034, | |
| "step": 13600 | |
| }, | |
| { | |
| "epoch": 17.68, | |
| "learning_rate": 0.00019641943734015346, | |
| "loss": 0.3062, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 17.68, | |
| "eval_cer": 0.3091941782037629, | |
| "eval_loss": 1.1049782037734985, | |
| "eval_runtime": 22.7193, | |
| "eval_samples_per_second": 22.756, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 18.18, | |
| "learning_rate": 0.00019335038363171354, | |
| "loss": 0.2861, | |
| "step": 14400 | |
| }, | |
| { | |
| "epoch": 18.18, | |
| "eval_cer": 0.30085197018104365, | |
| "eval_loss": 1.1057851314544678, | |
| "eval_runtime": 23.1091, | |
| "eval_samples_per_second": 22.372, | |
| "step": 14400 | |
| }, | |
| { | |
| "epoch": 18.69, | |
| "learning_rate": 0.00019028132992327364, | |
| "loss": 0.2669, | |
| "step": 14800 | |
| }, | |
| { | |
| "epoch": 18.69, | |
| "eval_cer": 0.30404685835995743, | |
| "eval_loss": 1.0846809148788452, | |
| "eval_runtime": 23.1356, | |
| "eval_samples_per_second": 22.347, | |
| "step": 14800 | |
| }, | |
| { | |
| "epoch": 19.19, | |
| "learning_rate": 0.00018721227621483374, | |
| "loss": 0.2599, | |
| "step": 15200 | |
| }, | |
| { | |
| "epoch": 19.19, | |
| "eval_cer": 0.30014199503017397, | |
| "eval_loss": 1.0948718786239624, | |
| "eval_runtime": 23.9281, | |
| "eval_samples_per_second": 21.606, | |
| "step": 15200 | |
| }, | |
| { | |
| "epoch": 19.7, | |
| "learning_rate": 0.00018414322250639385, | |
| "loss": 0.2485, | |
| "step": 15600 | |
| }, | |
| { | |
| "epoch": 19.7, | |
| "eval_cer": 0.30102946396876107, | |
| "eval_loss": 1.0906890630722046, | |
| "eval_runtime": 22.3527, | |
| "eval_samples_per_second": 23.129, | |
| "step": 15600 | |
| }, | |
| { | |
| "epoch": 20.2, | |
| "learning_rate": 0.00018107416879795392, | |
| "loss": 0.2512, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 20.2, | |
| "eval_cer": 0.306354277600284, | |
| "eval_loss": 1.145861268043518, | |
| "eval_runtime": 22.0419, | |
| "eval_samples_per_second": 23.455, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 20.71, | |
| "learning_rate": 0.00017800511508951405, | |
| "loss": 0.2443, | |
| "step": 16400 | |
| }, | |
| { | |
| "epoch": 20.71, | |
| "eval_cer": 0.3015619453319134, | |
| "eval_loss": 1.0991747379302979, | |
| "eval_runtime": 22.3023, | |
| "eval_samples_per_second": 23.182, | |
| "step": 16400 | |
| }, | |
| { | |
| "epoch": 21.21, | |
| "learning_rate": 0.00017493606138107415, | |
| "loss": 0.2259, | |
| "step": 16800 | |
| }, | |
| { | |
| "epoch": 21.21, | |
| "eval_cer": 0.30014199503017397, | |
| "eval_loss": 1.1134285926818848, | |
| "eval_runtime": 23.1155, | |
| "eval_samples_per_second": 22.366, | |
| "step": 16800 | |
| }, | |
| { | |
| "epoch": 21.72, | |
| "learning_rate": 0.00017186700767263426, | |
| "loss": 0.2203, | |
| "step": 17200 | |
| }, | |
| { | |
| "epoch": 21.72, | |
| "eval_cer": 0.3028044018459354, | |
| "eval_loss": 1.1040829420089722, | |
| "eval_runtime": 22.7175, | |
| "eval_samples_per_second": 22.758, | |
| "step": 17200 | |
| }, | |
| { | |
| "epoch": 22.22, | |
| "learning_rate": 0.00016879795396419436, | |
| "loss": 0.2178, | |
| "step": 17600 | |
| }, | |
| { | |
| "epoch": 22.22, | |
| "eval_cer": 0.2944621938232162, | |
| "eval_loss": 1.0857818126678467, | |
| "eval_runtime": 23.0266, | |
| "eval_samples_per_second": 22.452, | |
| "step": 17600 | |
| }, | |
| { | |
| "epoch": 22.73, | |
| "learning_rate": 0.00016572890025575446, | |
| "loss": 0.2008, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 22.73, | |
| "eval_cer": 0.29960951366702165, | |
| "eval_loss": 1.0990114212036133, | |
| "eval_runtime": 22.7824, | |
| "eval_samples_per_second": 22.693, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 23.23, | |
| "learning_rate": 0.00016265984654731456, | |
| "loss": 0.2037, | |
| "step": 18400 | |
| }, | |
| { | |
| "epoch": 23.23, | |
| "eval_cer": 0.2916222932197373, | |
| "eval_loss": 1.0890922546386719, | |
| "eval_runtime": 22.1133, | |
| "eval_samples_per_second": 23.38, | |
| "step": 18400 | |
| }, | |
| { | |
| "epoch": 23.74, | |
| "learning_rate": 0.0001595907928388747, | |
| "loss": 0.1885, | |
| "step": 18800 | |
| }, | |
| { | |
| "epoch": 23.74, | |
| "eval_cer": 0.2955271565495208, | |
| "eval_loss": 1.1051559448242188, | |
| "eval_runtime": 21.9892, | |
| "eval_samples_per_second": 23.512, | |
| "step": 18800 | |
| } | |
| ], | |
| "max_steps": 39600, | |
| "num_train_epochs": 50, | |
| "total_flos": 3.830606754314487e+19, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |