| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 4.9998887776665555, | |
| "global_step": 22475, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.11, | |
| "learning_rate": 2.982e-05, | |
| "loss": 10.9521, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "learning_rate": 2.9866994413407824e-05, | |
| "loss": 2.9894, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "eval_cer": 1.0, | |
| "eval_loss": 2.925715684890747, | |
| "eval_runtime": 1316.2549, | |
| "eval_samples_per_second": 11.314, | |
| "eval_steps_per_second": 1.415, | |
| "eval_wer": 1.0, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "learning_rate": 2.973291620111732e-05, | |
| "loss": 2.6003, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "learning_rate": 2.9598837988826818e-05, | |
| "loss": 0.7104, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "eval_cer": 0.045692850294096206, | |
| "eval_loss": 0.2129213511943817, | |
| "eval_runtime": 1315.3965, | |
| "eval_samples_per_second": 11.321, | |
| "eval_steps_per_second": 1.416, | |
| "eval_wer": 0.25382939389937936, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "learning_rate": 2.9464759776536317e-05, | |
| "loss": 0.3546, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "learning_rate": 2.933068156424581e-05, | |
| "loss": 0.2853, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "eval_cer": 0.027426083445944053, | |
| "eval_loss": 0.11085120588541031, | |
| "eval_runtime": 1321.34, | |
| "eval_samples_per_second": 11.27, | |
| "eval_steps_per_second": 1.409, | |
| "eval_wer": 0.15825135349267133, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "learning_rate": 2.9196603351955308e-05, | |
| "loss": 0.2563, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 0.89, | |
| "learning_rate": 2.9062525139664803e-05, | |
| "loss": 0.2327, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.89, | |
| "eval_cer": 0.02314500709307346, | |
| "eval_loss": 0.09091737866401672, | |
| "eval_runtime": 1319.6784, | |
| "eval_samples_per_second": 11.285, | |
| "eval_steps_per_second": 1.411, | |
| "eval_wer": 0.1320315594876535, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "learning_rate": 2.8928446927374302e-05, | |
| "loss": 0.2072, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 1.11, | |
| "learning_rate": 2.8794368715083797e-05, | |
| "loss": 0.1917, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 1.11, | |
| "eval_cer": 0.02055449330783939, | |
| "eval_loss": 0.07748183608055115, | |
| "eval_runtime": 1318.284, | |
| "eval_samples_per_second": 11.297, | |
| "eval_steps_per_second": 1.412, | |
| "eval_wer": 0.11881024692988247, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 1.22, | |
| "learning_rate": 2.8660290502793296e-05, | |
| "loss": 0.1853, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 1.33, | |
| "learning_rate": 2.8526212290502795e-05, | |
| "loss": 0.1803, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 1.33, | |
| "eval_cer": 0.018439187408532995, | |
| "eval_loss": 0.0698164626955986, | |
| "eval_runtime": 1319.3199, | |
| "eval_samples_per_second": 11.288, | |
| "eval_steps_per_second": 1.411, | |
| "eval_wer": 0.10547339231480259, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 1.45, | |
| "learning_rate": 2.839213407821229e-05, | |
| "loss": 0.1724, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 1.56, | |
| "learning_rate": 2.825805586592179e-05, | |
| "loss": 0.1661, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 1.56, | |
| "eval_cer": 0.01694487588523239, | |
| "eval_loss": 0.06446516513824463, | |
| "eval_runtime": 1319.092, | |
| "eval_samples_per_second": 11.29, | |
| "eval_steps_per_second": 1.412, | |
| "eval_wer": 0.09606496764822395, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 1.67, | |
| "learning_rate": 2.8123977653631284e-05, | |
| "loss": 0.161, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 1.78, | |
| "learning_rate": 2.7989899441340783e-05, | |
| "loss": 0.1635, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 1.78, | |
| "eval_cer": 0.017010760164402304, | |
| "eval_loss": 0.0639302060008049, | |
| "eval_runtime": 1319.0191, | |
| "eval_samples_per_second": 11.29, | |
| "eval_steps_per_second": 1.412, | |
| "eval_wer": 0.09642809982833751, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 1.89, | |
| "learning_rate": 2.785582122905028e-05, | |
| "loss": 0.1547, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "learning_rate": 2.7721743016759777e-05, | |
| "loss": 0.1555, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_cer": 0.015554297056795053, | |
| "eval_loss": 0.0592055507004261, | |
| "eval_runtime": 1320.0874, | |
| "eval_samples_per_second": 11.281, | |
| "eval_steps_per_second": 1.411, | |
| "eval_wer": 0.08808431268981909, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 2.11, | |
| "learning_rate": 2.7587664804469276e-05, | |
| "loss": 0.1405, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 2.22, | |
| "learning_rate": 2.745358659217877e-05, | |
| "loss": 0.1386, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 2.22, | |
| "eval_cer": 0.014650140459676017, | |
| "eval_loss": 0.055936481803655624, | |
| "eval_runtime": 1320.1961, | |
| "eval_samples_per_second": 11.28, | |
| "eval_steps_per_second": 1.41, | |
| "eval_wer": 0.08208437871385184, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 2.34, | |
| "learning_rate": 2.731950837988827e-05, | |
| "loss": 0.1353, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 2.45, | |
| "learning_rate": 2.7185430167597765e-05, | |
| "loss": 0.1338, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 2.45, | |
| "eval_cer": 0.014644533286980706, | |
| "eval_loss": 0.05477109178900719, | |
| "eval_runtime": 1321.7173, | |
| "eval_samples_per_second": 11.267, | |
| "eval_steps_per_second": 1.409, | |
| "eval_wer": 0.08314901624191205, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 2.56, | |
| "learning_rate": 2.7051620111731843e-05, | |
| "loss": 0.1345, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 2.67, | |
| "learning_rate": 2.691754189944134e-05, | |
| "loss": 0.1307, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 2.67, | |
| "eval_cer": 0.013712340826385111, | |
| "eval_loss": 0.0528746023774147, | |
| "eval_runtime": 1321.5984, | |
| "eval_samples_per_second": 11.268, | |
| "eval_steps_per_second": 1.409, | |
| "eval_wer": 0.07593589066420177, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 2.78, | |
| "learning_rate": 1.3624573378839592e-05, | |
| "loss": 0.1264, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 2.89, | |
| "learning_rate": 1.29419795221843e-05, | |
| "loss": 0.1297, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 2.89, | |
| "eval_cer": 0.013381517637361713, | |
| "eval_loss": 0.05041037127375603, | |
| "eval_runtime": 1303.0852, | |
| "eval_samples_per_second": 11.428, | |
| "eval_steps_per_second": 1.429, | |
| "eval_wer": 0.07453287996830846, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "learning_rate": 1.225938566552901e-05, | |
| "loss": 0.122, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 3.11, | |
| "learning_rate": 1.1578156996587031e-05, | |
| "loss": 0.1201, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 3.11, | |
| "eval_cer": 0.013103962588943778, | |
| "eval_loss": 0.04994073510169983, | |
| "eval_runtime": 1305.505, | |
| "eval_samples_per_second": 11.407, | |
| "eval_steps_per_second": 1.426, | |
| "eval_wer": 0.07336095338703288, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 3.23, | |
| "learning_rate": 1.089556313993174e-05, | |
| "loss": 0.1185, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 3.34, | |
| "learning_rate": 1.0214334470989761e-05, | |
| "loss": 0.1152, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 3.34, | |
| "eval_cer": 0.012759121468182099, | |
| "eval_loss": 0.0484052337706089, | |
| "eval_runtime": 1304.2248, | |
| "eval_samples_per_second": 11.418, | |
| "eval_steps_per_second": 1.428, | |
| "eval_wer": 0.07121517232272548, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 3.45, | |
| "learning_rate": 9.531740614334472e-06, | |
| "loss": 0.1165, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 3.56, | |
| "learning_rate": 8.849146757679181e-06, | |
| "loss": 0.1144, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 3.56, | |
| "eval_cer": 0.012485771799285646, | |
| "eval_loss": 0.04772350192070007, | |
| "eval_runtime": 1302.5642, | |
| "eval_samples_per_second": 11.433, | |
| "eval_steps_per_second": 1.429, | |
| "eval_wer": 0.06949029446718606, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 3.67, | |
| "learning_rate": 8.166552901023892e-06, | |
| "loss": 0.1129, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 3.78, | |
| "learning_rate": 7.483959044368601e-06, | |
| "loss": 0.1179, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 3.78, | |
| "eval_cer": 0.012212422130389193, | |
| "eval_loss": 0.04678424075245857, | |
| "eval_runtime": 1303.8755, | |
| "eval_samples_per_second": 11.421, | |
| "eval_steps_per_second": 1.428, | |
| "eval_wer": 0.0678809586689555, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 3.89, | |
| "learning_rate": 6.8040955631399314e-06, | |
| "loss": 0.1118, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "learning_rate": 6.1215017064846416e-06, | |
| "loss": 0.1112, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_cer": 0.012117100194568893, | |
| "eval_loss": 0.04677628353238106, | |
| "eval_runtime": 1306.4884, | |
| "eval_samples_per_second": 11.398, | |
| "eval_steps_per_second": 1.425, | |
| "eval_wer": 0.0676416215502443, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 4.12, | |
| "learning_rate": 5.438907849829352e-06, | |
| "loss": 0.1117, | |
| "step": 18500 | |
| }, | |
| { | |
| "epoch": 4.23, | |
| "learning_rate": 4.756313993174062e-06, | |
| "loss": 0.1141, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 4.23, | |
| "eval_cer": 0.012068037433484913, | |
| "eval_loss": 0.04616771265864372, | |
| "eval_runtime": 1302.5466, | |
| "eval_samples_per_second": 11.433, | |
| "eval_steps_per_second": 1.43, | |
| "eval_wer": 0.06683282714908227, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 4.34, | |
| "learning_rate": 4.073720136518771e-06, | |
| "loss": 0.1108, | |
| "step": 19500 | |
| }, | |
| { | |
| "epoch": 4.45, | |
| "learning_rate": 3.3911262798634816e-06, | |
| "loss": 0.1085, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 4.45, | |
| "eval_cer": 0.011937670668318913, | |
| "eval_loss": 0.04581034556031227, | |
| "eval_runtime": 1302.9709, | |
| "eval_samples_per_second": 11.429, | |
| "eval_steps_per_second": 1.429, | |
| "eval_wer": 0.06640367093622078, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 4.56, | |
| "learning_rate": 2.708532423208191e-06, | |
| "loss": 0.1132, | |
| "step": 20500 | |
| }, | |
| { | |
| "epoch": 4.67, | |
| "learning_rate": 2.025938566552901e-06, | |
| "loss": 0.105, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 4.67, | |
| "eval_cer": 0.01189281328675642, | |
| "eval_loss": 0.04555269330739975, | |
| "eval_runtime": 1307.8114, | |
| "eval_samples_per_second": 11.387, | |
| "eval_steps_per_second": 1.424, | |
| "eval_wer": 0.06604053875610723, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 4.78, | |
| "learning_rate": 1.3433447098976108e-06, | |
| "loss": 0.1123, | |
| "step": 21500 | |
| }, | |
| { | |
| "epoch": 4.89, | |
| "learning_rate": 6.621160409556314e-07, | |
| "loss": 0.1072, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 4.89, | |
| "eval_cer": 0.011854964871063065, | |
| "eval_loss": 0.04538210481405258, | |
| "eval_runtime": 1302.5621, | |
| "eval_samples_per_second": 11.433, | |
| "eval_steps_per_second": 1.429, | |
| "eval_wer": 0.06575168361283507, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "step": 22475, | |
| "total_flos": 1.360184451953694e+20, | |
| "train_loss": 0.0023267049651522526, | |
| "train_runtime": 2078.5155, | |
| "train_samples_per_second": 346.038, | |
| "train_steps_per_second": 10.813 | |
| } | |
| ], | |
| "max_steps": 22475, | |
| "num_train_epochs": 5, | |
| "total_flos": 1.360184451953694e+20, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |