breeze-listen-dsw-base-te / trainer_state.json
hanasim's picture
End of training
deed02c verified
{
"best_metric": 37.45436058603319,
"best_model_checkpoint": "/cosmos/home/sp-operator/ai/training/models/huggingface/scripts/../breeze-listen-dsw-base-te/checkpoint-800",
"epoch": 27.028,
"eval_steps": 200,
"global_step": 2000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.01,
"learning_rate": 5.0453611334320685e-06,
"loss": 1.6929,
"step": 25
},
{
"epoch": 0.03,
"learning_rate": 6.229195710491767e-06,
"loss": 1.4259,
"step": 50
},
{
"epoch": 1.0,
"learning_rate": 6.903829450223392e-06,
"loss": 0.9613,
"step": 75
},
{
"epoch": 1.01,
"learning_rate": 7.377725845391017e-06,
"loss": 0.5931,
"step": 100
},
{
"epoch": 1.03,
"learning_rate": 7.743343231239583e-06,
"loss": 0.4493,
"step": 125
},
{
"epoch": 2.0,
"learning_rate": 8.041073861170494e-06,
"loss": 0.3757,
"step": 150
},
{
"epoch": 2.02,
"learning_rate": 8.292222957399574e-06,
"loss": 0.3184,
"step": 175
},
{
"epoch": 2.03,
"learning_rate": 8.509413541357755e-06,
"loss": 0.2937,
"step": 200
},
{
"epoch": 2.03,
"eval_loss": 0.32373046875,
"eval_runtime": 1211.7719,
"eval_samples_per_second": 0.383,
"eval_steps_per_second": 0.024,
"eval_wer": 42.56135323750982,
"step": 200
},
{
"epoch": 3.0,
"learning_rate": 8.700744577655557e-06,
"loss": 0.2688,
"step": 225
},
{
"epoch": 3.02,
"learning_rate": 8.871723942761204e-06,
"loss": 0.2441,
"step": 250
},
{
"epoch": 3.03,
"learning_rate": 9.026267958246849e-06,
"loss": 0.2283,
"step": 275
},
{
"epoch": 4.01,
"learning_rate": 9.16726106663399e-06,
"loss": 0.2079,
"step": 300
},
{
"epoch": 4.02,
"learning_rate": 9.296889251455016e-06,
"loss": 0.1954,
"step": 325
},
{
"epoch": 4.03,
"learning_rate": 9.416848797368692e-06,
"loss": 0.1879,
"step": 350
},
{
"epoch": 5.01,
"learning_rate": 9.528482449516371e-06,
"loss": 0.1727,
"step": 375
},
{
"epoch": 5.02,
"learning_rate": 9.632871309784314e-06,
"loss": 0.1611,
"step": 400
},
{
"epoch": 5.02,
"eval_loss": 0.275634765625,
"eval_runtime": 1184.4646,
"eval_samples_per_second": 0.392,
"eval_steps_per_second": 0.024,
"eval_wer": 38.9148218329713,
"step": 400
},
{
"epoch": 5.03,
"learning_rate": 9.73089868785391e-06,
"loss": 0.1524,
"step": 425
},
{
"epoch": 6.01,
"learning_rate": 9.823295589572114e-06,
"loss": 0.1468,
"step": 450
},
{
"epoch": 6.02,
"learning_rate": 9.910673836465484e-06,
"loss": 0.1323,
"step": 475
},
{
"epoch": 6.03,
"learning_rate": 9.993550644973805e-06,
"loss": 0.1289,
"step": 500
},
{
"epoch": 7.01,
"learning_rate": 9.853333333333334e-06,
"loss": 0.1159,
"step": 525
},
{
"epoch": 7.02,
"learning_rate": 9.686666666666668e-06,
"loss": 0.1075,
"step": 550
},
{
"epoch": 7.04,
"learning_rate": 9.52e-06,
"loss": 0.1032,
"step": 575
},
{
"epoch": 8.01,
"learning_rate": 9.353333333333334e-06,
"loss": 0.0889,
"step": 600
},
{
"epoch": 8.01,
"eval_loss": 0.29296875,
"eval_runtime": 1184.8459,
"eval_samples_per_second": 0.392,
"eval_steps_per_second": 0.024,
"eval_wer": 38.11064380459398,
"step": 600
},
{
"epoch": 8.02,
"learning_rate": 9.186666666666666e-06,
"loss": 0.0861,
"step": 625
},
{
"epoch": 9.0,
"learning_rate": 9.020000000000002e-06,
"loss": 0.0806,
"step": 650
},
{
"epoch": 9.01,
"learning_rate": 8.853333333333334e-06,
"loss": 0.0681,
"step": 675
},
{
"epoch": 9.03,
"learning_rate": 8.686666666666668e-06,
"loss": 0.0708,
"step": 700
},
{
"epoch": 10.0,
"learning_rate": 8.52e-06,
"loss": 0.0626,
"step": 725
},
{
"epoch": 10.02,
"learning_rate": 8.353333333333335e-06,
"loss": 0.0552,
"step": 750
},
{
"epoch": 10.03,
"learning_rate": 8.186666666666667e-06,
"loss": 0.0517,
"step": 775
},
{
"epoch": 11.0,
"learning_rate": 8.020000000000001e-06,
"loss": 0.0456,
"step": 800
},
{
"epoch": 11.0,
"eval_loss": 0.337158203125,
"eval_runtime": 1186.6186,
"eval_samples_per_second": 0.391,
"eval_steps_per_second": 0.024,
"eval_wer": 37.45436058603319,
"step": 800
},
{
"epoch": 11.02,
"learning_rate": 7.853333333333333e-06,
"loss": 0.0404,
"step": 825
},
{
"epoch": 11.03,
"learning_rate": 7.686666666666667e-06,
"loss": 0.0407,
"step": 850
},
{
"epoch": 12.01,
"learning_rate": 7.520000000000001e-06,
"loss": 0.0369,
"step": 875
},
{
"epoch": 12.02,
"learning_rate": 7.353333333333334e-06,
"loss": 0.0303,
"step": 900
},
{
"epoch": 12.03,
"learning_rate": 7.186666666666668e-06,
"loss": 0.0289,
"step": 925
},
{
"epoch": 13.01,
"learning_rate": 7.0200000000000006e-06,
"loss": 0.0268,
"step": 950
},
{
"epoch": 13.02,
"learning_rate": 6.853333333333334e-06,
"loss": 0.0238,
"step": 975
},
{
"epoch": 13.03,
"learning_rate": 6.6866666666666665e-06,
"loss": 0.0229,
"step": 1000
},
{
"epoch": 13.03,
"eval_loss": 0.398193359375,
"eval_runtime": 1184.9328,
"eval_samples_per_second": 0.392,
"eval_steps_per_second": 0.024,
"eval_wer": 37.92577529232333,
"step": 1000
},
{
"epoch": 14.01,
"learning_rate": 6.533333333333334e-06,
"loss": 0.0195,
"step": 1025
},
{
"epoch": 14.02,
"learning_rate": 6.366666666666668e-06,
"loss": 0.0177,
"step": 1050
},
{
"epoch": 14.03,
"learning_rate": 6.200000000000001e-06,
"loss": 0.0176,
"step": 1075
},
{
"epoch": 15.01,
"learning_rate": 6.033333333333335e-06,
"loss": 0.0149,
"step": 1100
},
{
"epoch": 15.02,
"learning_rate": 5.8666666666666675e-06,
"loss": 0.014,
"step": 1125
},
{
"epoch": 15.04,
"learning_rate": 5.7e-06,
"loss": 0.0124,
"step": 1150
},
{
"epoch": 16.01,
"learning_rate": 5.533333333333334e-06,
"loss": 0.0095,
"step": 1175
},
{
"epoch": 16.02,
"learning_rate": 5.366666666666666e-06,
"loss": 0.0103,
"step": 1200
},
{
"epoch": 16.02,
"eval_loss": 0.447265625,
"eval_runtime": 1183.5656,
"eval_samples_per_second": 0.392,
"eval_steps_per_second": 0.025,
"eval_wer": 38.26778204002403,
"step": 1200
},
{
"epoch": 17.0,
"learning_rate": 5.2e-06,
"loss": 0.01,
"step": 1225
},
{
"epoch": 17.01,
"learning_rate": 5.033333333333333e-06,
"loss": 0.0081,
"step": 1250
},
{
"epoch": 17.03,
"learning_rate": 4.866666666666667e-06,
"loss": 0.0078,
"step": 1275
},
{
"epoch": 18.0,
"learning_rate": 4.7e-06,
"loss": 0.0073,
"step": 1300
},
{
"epoch": 18.01,
"learning_rate": 4.533333333333334e-06,
"loss": 0.0053,
"step": 1325
},
{
"epoch": 18.03,
"learning_rate": 4.366666666666667e-06,
"loss": 0.0059,
"step": 1350
},
{
"epoch": 19.0,
"learning_rate": 4.2000000000000004e-06,
"loss": 0.0051,
"step": 1375
},
{
"epoch": 19.02,
"learning_rate": 4.033333333333333e-06,
"loss": 0.0042,
"step": 1400
},
{
"epoch": 19.02,
"eval_loss": 0.483642578125,
"eval_runtime": 1188.6164,
"eval_samples_per_second": 0.39,
"eval_steps_per_second": 0.024,
"eval_wer": 37.89804501548274,
"step": 1400
},
{
"epoch": 19.03,
"learning_rate": 3.866666666666667e-06,
"loss": 0.0043,
"step": 1425
},
{
"epoch": 20.0,
"learning_rate": 3.7e-06,
"loss": 0.004,
"step": 1450
},
{
"epoch": 20.02,
"learning_rate": 3.5333333333333335e-06,
"loss": 0.0034,
"step": 1475
},
{
"epoch": 20.03,
"learning_rate": 3.366666666666667e-06,
"loss": 0.003,
"step": 1500
},
{
"epoch": 21.01,
"learning_rate": 3.2000000000000003e-06,
"loss": 0.0032,
"step": 1525
},
{
"epoch": 21.02,
"learning_rate": 3.0333333333333337e-06,
"loss": 0.0027,
"step": 1550
},
{
"epoch": 21.03,
"learning_rate": 2.866666666666667e-06,
"loss": 0.0029,
"step": 1575
},
{
"epoch": 22.01,
"learning_rate": 2.7000000000000004e-06,
"loss": 0.0025,
"step": 1600
},
{
"epoch": 22.01,
"eval_loss": 0.50830078125,
"eval_runtime": 1190.0425,
"eval_samples_per_second": 0.39,
"eval_steps_per_second": 0.024,
"eval_wer": 37.731663354439156,
"step": 1600
},
{
"epoch": 22.02,
"learning_rate": 2.5333333333333338e-06,
"loss": 0.0023,
"step": 1625
},
{
"epoch": 22.03,
"learning_rate": 2.3666666666666667e-06,
"loss": 0.0024,
"step": 1650
},
{
"epoch": 23.01,
"learning_rate": 2.2e-06,
"loss": 0.0021,
"step": 1675
},
{
"epoch": 23.02,
"learning_rate": 2.0333333333333335e-06,
"loss": 0.0022,
"step": 1700
},
{
"epoch": 23.03,
"learning_rate": 1.8666666666666669e-06,
"loss": 0.0022,
"step": 1725
},
{
"epoch": 24.01,
"learning_rate": 1.7000000000000002e-06,
"loss": 0.0021,
"step": 1750
},
{
"epoch": 24.02,
"learning_rate": 1.5333333333333334e-06,
"loss": 0.0018,
"step": 1775
},
{
"epoch": 24.04,
"learning_rate": 1.3666666666666668e-06,
"loss": 0.002,
"step": 1800
},
{
"epoch": 24.04,
"eval_loss": 0.52197265625,
"eval_runtime": 1185.5264,
"eval_samples_per_second": 0.391,
"eval_steps_per_second": 0.024,
"eval_wer": 37.80098904654065,
"step": 1800
},
{
"epoch": 25.01,
"learning_rate": 1.2000000000000002e-06,
"loss": 0.0018,
"step": 1825
},
{
"epoch": 25.02,
"learning_rate": 1.0333333333333333e-06,
"loss": 0.002,
"step": 1850
},
{
"epoch": 26.0,
"learning_rate": 8.666666666666668e-07,
"loss": 0.0019,
"step": 1875
},
{
"epoch": 26.01,
"learning_rate": 7.000000000000001e-07,
"loss": 0.0018,
"step": 1900
},
{
"epoch": 26.03,
"learning_rate": 5.333333333333335e-07,
"loss": 0.0018,
"step": 1925
},
{
"epoch": 27.0,
"learning_rate": 3.666666666666667e-07,
"loss": 0.0018,
"step": 1950
},
{
"epoch": 27.02,
"learning_rate": 2.0000000000000002e-07,
"loss": 0.0017,
"step": 1975
},
{
"epoch": 27.03,
"learning_rate": 3.333333333333334e-08,
"loss": 0.0018,
"step": 2000
},
{
"epoch": 27.03,
"eval_loss": 0.52685546875,
"eval_runtime": 1188.9989,
"eval_samples_per_second": 0.39,
"eval_steps_per_second": 0.024,
"eval_wer": 37.9026667282895,
"step": 2000
},
{
"epoch": 27.03,
"step": 2000,
"total_flos": 4.1020056982698066e+18,
"train_loss": 0.12189112710952758,
"train_runtime": 77912.2113,
"train_samples_per_second": 0.821,
"train_steps_per_second": 0.026
}
],
"logging_steps": 25,
"max_steps": 2000,
"num_input_tokens_seen": 0,
"num_train_epochs": 9223372036854775807,
"save_steps": 200,
"total_flos": 4.1020056982698066e+18,
"train_batch_size": 32,
"trial_name": null,
"trial_params": null
}