| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 5.999854956849663, | |
| "eval_steps": 4000, | |
| "global_step": 124098, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.19, | |
| "learning_rate": 9.756372971128695e-05, | |
| "loss": 0.7054, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "eval_cer": 0.02274243901983751, | |
| "eval_loss": 0.10111288726329803, | |
| "eval_runtime": 66.987, | |
| "eval_samples_per_second": 61.131, | |
| "eval_steps_per_second": 3.822, | |
| "eval_wer": 0.08709714849302352, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "learning_rate": 9.431509853937514e-05, | |
| "loss": 0.0856, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "eval_cer": 0.020731155298169313, | |
| "eval_loss": 0.09945787489414215, | |
| "eval_runtime": 67.6612, | |
| "eval_samples_per_second": 60.522, | |
| "eval_steps_per_second": 3.784, | |
| "eval_wer": 0.07468163490267746, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "learning_rate": 9.106727972834653e-05, | |
| "loss": 0.075, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "eval_cer": 0.018470211068126987, | |
| "eval_loss": 0.08679623156785965, | |
| "eval_runtime": 72.7557, | |
| "eval_samples_per_second": 56.284, | |
| "eval_steps_per_second": 3.519, | |
| "eval_wer": 0.06474632320946828, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "learning_rate": 8.781864855643472e-05, | |
| "loss": 0.0694, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "eval_cer": 0.01828354900579119, | |
| "eval_loss": 0.08532032370567322, | |
| "eval_runtime": 71.0749, | |
| "eval_samples_per_second": 57.615, | |
| "eval_steps_per_second": 3.602, | |
| "eval_wer": 0.06190351869579091, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "learning_rate": 8.45708297454061e-05, | |
| "loss": 0.0658, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "eval_cer": 0.017056245945933333, | |
| "eval_loss": 0.0778348445892334, | |
| "eval_runtime": 68.8966, | |
| "eval_samples_per_second": 59.437, | |
| "eval_steps_per_second": 3.716, | |
| "eval_wer": 0.057262205204072755, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 1.16, | |
| "learning_rate": 8.13230109343775e-05, | |
| "loss": 0.0589, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 1.16, | |
| "eval_cer": 0.016645589408794584, | |
| "eval_loss": 0.08211962133646011, | |
| "eval_runtime": 71.5048, | |
| "eval_samples_per_second": 57.269, | |
| "eval_steps_per_second": 3.58, | |
| "eval_wer": 0.05463696226031967, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 1.35, | |
| "learning_rate": 7.807437976246568e-05, | |
| "loss": 0.0572, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 1.35, | |
| "eval_cer": 0.0169955807756742, | |
| "eval_loss": 0.08269500732421875, | |
| "eval_runtime": 64.9267, | |
| "eval_samples_per_second": 63.071, | |
| "eval_steps_per_second": 3.943, | |
| "eval_wer": 0.05579729063324921, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 1.55, | |
| "learning_rate": 7.482656095143708e-05, | |
| "loss": 0.0551, | |
| "step": 32000 | |
| }, | |
| { | |
| "epoch": 1.55, | |
| "eval_cer": 0.016853250953143155, | |
| "eval_loss": 0.08304612338542938, | |
| "eval_runtime": 67.5544, | |
| "eval_samples_per_second": 60.618, | |
| "eval_steps_per_second": 3.79, | |
| "eval_wer": 0.05333159284077394, | |
| "step": 32000 | |
| }, | |
| { | |
| "epoch": 1.74, | |
| "learning_rate": 7.157874214040847e-05, | |
| "loss": 0.054, | |
| "step": 36000 | |
| }, | |
| { | |
| "epoch": 1.74, | |
| "eval_cer": 0.016157934770942316, | |
| "eval_loss": 0.0788031816482544, | |
| "eval_runtime": 68.7321, | |
| "eval_samples_per_second": 59.579, | |
| "eval_steps_per_second": 3.725, | |
| "eval_wer": 0.05122849766483915, | |
| "step": 36000 | |
| }, | |
| { | |
| "epoch": 1.93, | |
| "learning_rate": 6.832929860761345e-05, | |
| "loss": 0.0524, | |
| "step": 40000 | |
| }, | |
| { | |
| "epoch": 1.93, | |
| "eval_cer": 0.015562949447246968, | |
| "eval_loss": 0.07834739238023758, | |
| "eval_runtime": 66.4802, | |
| "eval_samples_per_second": 61.597, | |
| "eval_steps_per_second": 3.851, | |
| "eval_wer": 0.04893684912830331, | |
| "step": 40000 | |
| }, | |
| { | |
| "epoch": 2.13, | |
| "learning_rate": 6.508147979658484e-05, | |
| "loss": 0.048, | |
| "step": 44000 | |
| }, | |
| { | |
| "epoch": 2.13, | |
| "eval_cer": 0.016024938051528062, | |
| "eval_loss": 0.08605939149856567, | |
| "eval_runtime": 68.4968, | |
| "eval_samples_per_second": 59.784, | |
| "eval_steps_per_second": 3.737, | |
| "eval_wer": 0.04918341890755084, | |
| "step": 44000 | |
| }, | |
| { | |
| "epoch": 2.32, | |
| "learning_rate": 6.183284862467303e-05, | |
| "loss": 0.046, | |
| "step": 48000 | |
| }, | |
| { | |
| "epoch": 2.32, | |
| "eval_cer": 0.015420619624715923, | |
| "eval_loss": 0.07625599950551987, | |
| "eval_runtime": 73.2954, | |
| "eval_samples_per_second": 55.87, | |
| "eval_steps_per_second": 3.493, | |
| "eval_wer": 0.049400980477475126, | |
| "step": 48000 | |
| }, | |
| { | |
| "epoch": 2.51, | |
| "learning_rate": 5.8584217452761215e-05, | |
| "loss": 0.0456, | |
| "step": 52000 | |
| }, | |
| { | |
| "epoch": 2.51, | |
| "eval_cer": 0.015303955835756052, | |
| "eval_loss": 0.08352649956941605, | |
| "eval_runtime": 71.5001, | |
| "eval_samples_per_second": 57.273, | |
| "eval_steps_per_second": 3.58, | |
| "eval_wer": 0.047094827836277664, | |
| "step": 52000 | |
| }, | |
| { | |
| "epoch": 2.71, | |
| "learning_rate": 5.53363986417326e-05, | |
| "loss": 0.0439, | |
| "step": 56000 | |
| }, | |
| { | |
| "epoch": 2.71, | |
| "eval_cer": 0.015152292910108217, | |
| "eval_loss": 0.07897598296403885, | |
| "eval_runtime": 71.5932, | |
| "eval_samples_per_second": 57.198, | |
| "eval_steps_per_second": 3.576, | |
| "eval_wer": 0.04686276216169176, | |
| "step": 56000 | |
| }, | |
| { | |
| "epoch": 2.9, | |
| "learning_rate": 5.2088579830704e-05, | |
| "loss": 0.0436, | |
| "step": 60000 | |
| }, | |
| { | |
| "epoch": 2.9, | |
| "eval_cer": 0.0155349501378966, | |
| "eval_loss": 0.08321597427129745, | |
| "eval_runtime": 69.796, | |
| "eval_samples_per_second": 58.671, | |
| "eval_steps_per_second": 3.668, | |
| "eval_wer": 0.047196356568909, | |
| "step": 60000 | |
| }, | |
| { | |
| "epoch": 3.09, | |
| "learning_rate": 4.8840761019675384e-05, | |
| "loss": 0.0406, | |
| "step": 64000 | |
| }, | |
| { | |
| "epoch": 3.09, | |
| "eval_cer": 0.014837300679916562, | |
| "eval_loss": 0.08103086799383163, | |
| "eval_runtime": 69.1811, | |
| "eval_samples_per_second": 59.192, | |
| "eval_steps_per_second": 3.7, | |
| "eval_wer": 0.04416499869463058, | |
| "step": 64000 | |
| }, | |
| { | |
| "epoch": 3.29, | |
| "learning_rate": 4.559212984776357e-05, | |
| "loss": 0.0386, | |
| "step": 68000 | |
| }, | |
| { | |
| "epoch": 3.29, | |
| "eval_cer": 0.014624972584009594, | |
| "eval_loss": 0.08100830018520355, | |
| "eval_runtime": 72.8594, | |
| "eval_samples_per_second": 56.204, | |
| "eval_steps_per_second": 3.514, | |
| "eval_wer": 0.043628346822150665, | |
| "step": 68000 | |
| }, | |
| { | |
| "epoch": 3.48, | |
| "learning_rate": 4.234349867585176e-05, | |
| "loss": 0.038, | |
| "step": 72000 | |
| }, | |
| { | |
| "epoch": 3.48, | |
| "eval_cer": 0.014267981389792385, | |
| "eval_loss": 0.07782719284296036, | |
| "eval_runtime": 73.706, | |
| "eval_samples_per_second": 55.559, | |
| "eval_steps_per_second": 3.473, | |
| "eval_wer": 0.04301917442636266, | |
| "step": 72000 | |
| }, | |
| { | |
| "epoch": 3.67, | |
| "learning_rate": 3.909567986482315e-05, | |
| "loss": 0.0373, | |
| "step": 76000 | |
| }, | |
| { | |
| "epoch": 3.67, | |
| "eval_cer": 0.014431310694336206, | |
| "eval_loss": 0.07849407941102982, | |
| "eval_runtime": 74.4035, | |
| "eval_samples_per_second": 55.038, | |
| "eval_steps_per_second": 3.441, | |
| "eval_wer": 0.042975662112377806, | |
| "step": 76000 | |
| }, | |
| { | |
| "epoch": 3.87, | |
| "learning_rate": 3.584786105379454e-05, | |
| "loss": 0.0363, | |
| "step": 80000 | |
| }, | |
| { | |
| "epoch": 3.87, | |
| "eval_cer": 0.014375312075635468, | |
| "eval_loss": 0.07884030044078827, | |
| "eval_runtime": 68.8676, | |
| "eval_samples_per_second": 59.462, | |
| "eval_steps_per_second": 3.717, | |
| "eval_wer": 0.042119919937342265, | |
| "step": 80000 | |
| }, | |
| { | |
| "epoch": 4.06, | |
| "learning_rate": 3.259922988188273e-05, | |
| "loss": 0.0348, | |
| "step": 84000 | |
| }, | |
| { | |
| "epoch": 4.06, | |
| "eval_cer": 0.01435431259362269, | |
| "eval_loss": 0.08232194930315018, | |
| "eval_runtime": 67.1145, | |
| "eval_samples_per_second": 61.015, | |
| "eval_steps_per_second": 3.814, | |
| "eval_wer": 0.04232297740260494, | |
| "step": 84000 | |
| }, | |
| { | |
| "epoch": 4.25, | |
| "learning_rate": 2.935059870997092e-05, | |
| "loss": 0.0323, | |
| "step": 88000 | |
| }, | |
| { | |
| "epoch": 4.25, | |
| "eval_cer": 0.014316980181155532, | |
| "eval_loss": 0.08194055408239365, | |
| "eval_runtime": 76.1231, | |
| "eval_samples_per_second": 53.794, | |
| "eval_steps_per_second": 3.363, | |
| "eval_wer": 0.04068401357584196, | |
| "step": 88000 | |
| }, | |
| { | |
| "epoch": 4.45, | |
| "learning_rate": 2.610277989894231e-05, | |
| "loss": 0.0319, | |
| "step": 92000 | |
| }, | |
| { | |
| "epoch": 4.45, | |
| "eval_cer": 0.014167650531286895, | |
| "eval_loss": 0.08085062354803085, | |
| "eval_runtime": 74.2762, | |
| "eval_samples_per_second": 55.132, | |
| "eval_steps_per_second": 3.447, | |
| "eval_wer": 0.041032112087720826, | |
| "step": 92000 | |
| }, | |
| { | |
| "epoch": 4.64, | |
| "learning_rate": 2.2854148727030496e-05, | |
| "loss": 0.0314, | |
| "step": 96000 | |
| }, | |
| { | |
| "epoch": 4.64, | |
| "eval_cer": 0.01383399209486166, | |
| "eval_loss": 0.08210451155900955, | |
| "eval_runtime": 73.1195, | |
| "eval_samples_per_second": 56.004, | |
| "eval_steps_per_second": 3.501, | |
| "eval_wer": 0.040045832970730715, | |
| "step": 96000 | |
| }, | |
| { | |
| "epoch": 4.83, | |
| "learning_rate": 1.9606329916001885e-05, | |
| "loss": 0.0306, | |
| "step": 100000 | |
| }, | |
| { | |
| "epoch": 4.83, | |
| "eval_cer": 0.013700995375447406, | |
| "eval_loss": 0.08130906522274017, | |
| "eval_runtime": 70.4204, | |
| "eval_samples_per_second": 58.151, | |
| "eval_steps_per_second": 3.635, | |
| "eval_wer": 0.03885649638847794, | |
| "step": 100000 | |
| }, | |
| { | |
| "epoch": 5.03, | |
| "learning_rate": 1.6358511104973274e-05, | |
| "loss": 0.0295, | |
| "step": 104000 | |
| }, | |
| { | |
| "epoch": 5.03, | |
| "eval_cer": 0.013115343154868847, | |
| "eval_loss": 0.0819702297449112, | |
| "eval_runtime": 67.1904, | |
| "eval_samples_per_second": 60.946, | |
| "eval_steps_per_second": 3.81, | |
| "eval_wer": 0.037710672120210016, | |
| "step": 104000 | |
| }, | |
| { | |
| "epoch": 5.22, | |
| "learning_rate": 1.3109879933061465e-05, | |
| "loss": 0.0275, | |
| "step": 108000 | |
| }, | |
| { | |
| "epoch": 5.22, | |
| "eval_cer": 0.013700995375447406, | |
| "eval_loss": 0.08659966289997101, | |
| "eval_runtime": 75.5877, | |
| "eval_samples_per_second": 54.176, | |
| "eval_steps_per_second": 3.387, | |
| "eval_wer": 0.037826704957502975, | |
| "step": 108000 | |
| }, | |
| { | |
| "epoch": 5.41, | |
| "learning_rate": 9.862061122032852e-06, | |
| "loss": 0.0267, | |
| "step": 112000 | |
| }, | |
| { | |
| "epoch": 5.41, | |
| "eval_cer": 0.013351004008567788, | |
| "eval_loss": 0.08306384831666946, | |
| "eval_runtime": 71.1622, | |
| "eval_samples_per_second": 57.545, | |
| "eval_steps_per_second": 3.597, | |
| "eval_wer": 0.03755112696893221, | |
| "step": 112000 | |
| }, | |
| { | |
| "epoch": 5.61, | |
| "learning_rate": 6.6134299501210415e-06, | |
| "loss": 0.0264, | |
| "step": 116000 | |
| }, | |
| { | |
| "epoch": 5.61, | |
| "eval_cer": 0.013194674531361559, | |
| "eval_loss": 0.08454854041337967, | |
| "eval_runtime": 69.2173, | |
| "eval_samples_per_second": 59.162, | |
| "eval_steps_per_second": 3.698, | |
| "eval_wer": 0.03691294636382096, | |
| "step": 116000 | |
| }, | |
| { | |
| "epoch": 5.8, | |
| "learning_rate": 3.364798778209232e-06, | |
| "loss": 0.0258, | |
| "step": 120000 | |
| }, | |
| { | |
| "epoch": 5.8, | |
| "eval_cer": 0.013325337974996616, | |
| "eval_loss": 0.08589179813861847, | |
| "eval_runtime": 73.5677, | |
| "eval_samples_per_second": 55.663, | |
| "eval_steps_per_second": 3.48, | |
| "eval_wer": 0.037028979201113914, | |
| "step": 120000 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "learning_rate": 1.1616760629742156e-07, | |
| "loss": 0.0254, | |
| "step": 124000 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_cer": 0.013243673322724706, | |
| "eval_loss": 0.0846036821603775, | |
| "eval_runtime": 67.6027, | |
| "eval_samples_per_second": 60.575, | |
| "eval_steps_per_second": 3.787, | |
| "eval_wer": 0.03666637658457343, | |
| "step": 124000 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "step": 124098, | |
| "total_flos": 8.637354948645209e+20, | |
| "train_loss": 0.0650978993577199, | |
| "train_runtime": 137826.0599, | |
| "train_samples_per_second": 28.813, | |
| "train_steps_per_second": 0.9 | |
| } | |
| ], | |
| "logging_steps": 4000, | |
| "max_steps": 124098, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 6, | |
| "save_steps": 500, | |
| "total_flos": 8.637354948645209e+20, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |