warmestman's picture
Training in progress, step 4000, checkpoint
2460e8b verified
raw
history blame
20.5 kB
{
"best_metric": 38.127715747208626,
"best_model_checkpoint": "warmestman/whisper-large-v3-mn-cv-fleurs/checkpoint-4000",
"epoch": 23.952095808383234,
"eval_steps": 1000,
"global_step": 4000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.15,
"learning_rate": 5e-06,
"loss": 1.1872,
"step": 25
},
{
"epoch": 0.3,
"learning_rate": 1e-05,
"loss": 0.7033,
"step": 50
},
{
"epoch": 0.45,
"learning_rate": 1.5e-05,
"loss": 0.4953,
"step": 75
},
{
"epoch": 0.6,
"learning_rate": 2e-05,
"loss": 0.3952,
"step": 100
},
{
"epoch": 0.75,
"learning_rate": 2.5e-05,
"loss": 0.3674,
"step": 125
},
{
"epoch": 0.9,
"learning_rate": 3e-05,
"loss": 0.3351,
"step": 150
},
{
"epoch": 1.05,
"learning_rate": 3.5e-05,
"loss": 0.2835,
"step": 175
},
{
"epoch": 1.2,
"learning_rate": 4e-05,
"loss": 0.2188,
"step": 200
},
{
"epoch": 1.35,
"learning_rate": 4.5e-05,
"loss": 0.2255,
"step": 225
},
{
"epoch": 1.5,
"learning_rate": 5e-05,
"loss": 0.22,
"step": 250
},
{
"epoch": 1.65,
"learning_rate": 5.500000000000001e-05,
"loss": 0.236,
"step": 275
},
{
"epoch": 1.8,
"learning_rate": 6e-05,
"loss": 0.2212,
"step": 300
},
{
"epoch": 1.95,
"learning_rate": 6.500000000000001e-05,
"loss": 0.2182,
"step": 325
},
{
"epoch": 2.1,
"learning_rate": 7e-05,
"loss": 0.1616,
"step": 350
},
{
"epoch": 2.25,
"learning_rate": 7.500000000000001e-05,
"loss": 0.1583,
"step": 375
},
{
"epoch": 2.4,
"learning_rate": 8e-05,
"loss": 0.1514,
"step": 400
},
{
"epoch": 2.54,
"learning_rate": 8.5e-05,
"loss": 0.1729,
"step": 425
},
{
"epoch": 2.69,
"learning_rate": 9e-05,
"loss": 0.1626,
"step": 450
},
{
"epoch": 2.84,
"learning_rate": 9.5e-05,
"loss": 0.1807,
"step": 475
},
{
"epoch": 2.99,
"learning_rate": 0.0001,
"loss": 0.1821,
"step": 500
},
{
"epoch": 3.14,
"learning_rate": 9.973684210526316e-05,
"loss": 0.1437,
"step": 525
},
{
"epoch": 3.29,
"learning_rate": 9.947368421052632e-05,
"loss": 0.1333,
"step": 550
},
{
"epoch": 3.44,
"learning_rate": 9.921052631578947e-05,
"loss": 0.1411,
"step": 575
},
{
"epoch": 3.59,
"learning_rate": 9.894736842105263e-05,
"loss": 0.1341,
"step": 600
},
{
"epoch": 3.74,
"learning_rate": 9.868421052631579e-05,
"loss": 0.1371,
"step": 625
},
{
"epoch": 3.89,
"learning_rate": 9.842105263157894e-05,
"loss": 0.1365,
"step": 650
},
{
"epoch": 4.04,
"learning_rate": 9.815789473684211e-05,
"loss": 0.118,
"step": 675
},
{
"epoch": 4.19,
"learning_rate": 9.789473684210527e-05,
"loss": 0.0827,
"step": 700
},
{
"epoch": 4.34,
"learning_rate": 9.763157894736843e-05,
"loss": 0.0879,
"step": 725
},
{
"epoch": 4.49,
"learning_rate": 9.736842105263158e-05,
"loss": 0.0937,
"step": 750
},
{
"epoch": 4.64,
"learning_rate": 9.710526315789474e-05,
"loss": 0.0887,
"step": 775
},
{
"epoch": 4.79,
"learning_rate": 9.68421052631579e-05,
"loss": 0.0946,
"step": 800
},
{
"epoch": 4.94,
"learning_rate": 9.657894736842105e-05,
"loss": 0.1032,
"step": 825
},
{
"epoch": 5.09,
"learning_rate": 9.631578947368421e-05,
"loss": 0.077,
"step": 850
},
{
"epoch": 5.24,
"learning_rate": 9.605263157894737e-05,
"loss": 0.0592,
"step": 875
},
{
"epoch": 5.39,
"learning_rate": 9.578947368421052e-05,
"loss": 0.0664,
"step": 900
},
{
"epoch": 5.54,
"learning_rate": 9.552631578947369e-05,
"loss": 0.066,
"step": 925
},
{
"epoch": 5.69,
"learning_rate": 9.526315789473685e-05,
"loss": 0.0641,
"step": 950
},
{
"epoch": 5.84,
"learning_rate": 9.5e-05,
"loss": 0.0662,
"step": 975
},
{
"epoch": 5.99,
"learning_rate": 9.473684210526316e-05,
"loss": 0.0691,
"step": 1000
},
{
"epoch": 5.99,
"eval_loss": 0.45974645018577576,
"eval_runtime": 585.199,
"eval_samples_per_second": 0.716,
"eval_steps_per_second": 0.091,
"eval_wer": 41.504867719047354,
"step": 1000
},
{
"epoch": 6.14,
"learning_rate": 9.447368421052633e-05,
"loss": 0.0455,
"step": 1025
},
{
"epoch": 6.29,
"learning_rate": 9.421052631578949e-05,
"loss": 0.0471,
"step": 1050
},
{
"epoch": 6.44,
"learning_rate": 9.394736842105264e-05,
"loss": 0.0471,
"step": 1075
},
{
"epoch": 6.59,
"learning_rate": 9.36842105263158e-05,
"loss": 0.0465,
"step": 1100
},
{
"epoch": 6.74,
"learning_rate": 9.342105263157896e-05,
"loss": 0.0486,
"step": 1125
},
{
"epoch": 6.89,
"learning_rate": 9.315789473684211e-05,
"loss": 0.044,
"step": 1150
},
{
"epoch": 7.04,
"learning_rate": 9.289473684210527e-05,
"loss": 0.0421,
"step": 1175
},
{
"epoch": 7.19,
"learning_rate": 9.263157894736843e-05,
"loss": 0.0333,
"step": 1200
},
{
"epoch": 7.34,
"learning_rate": 9.236842105263158e-05,
"loss": 0.0323,
"step": 1225
},
{
"epoch": 7.49,
"learning_rate": 9.210526315789474e-05,
"loss": 0.0289,
"step": 1250
},
{
"epoch": 7.63,
"learning_rate": 9.18421052631579e-05,
"loss": 0.0339,
"step": 1275
},
{
"epoch": 7.78,
"learning_rate": 9.157894736842105e-05,
"loss": 0.0332,
"step": 1300
},
{
"epoch": 7.93,
"learning_rate": 9.131578947368421e-05,
"loss": 0.0701,
"step": 1325
},
{
"epoch": 8.08,
"learning_rate": 9.105263157894738e-05,
"loss": 0.0601,
"step": 1350
},
{
"epoch": 8.23,
"learning_rate": 9.078947368421054e-05,
"loss": 0.0285,
"step": 1375
},
{
"epoch": 8.38,
"learning_rate": 9.052631578947369e-05,
"loss": 0.0258,
"step": 1400
},
{
"epoch": 8.53,
"learning_rate": 9.026315789473685e-05,
"loss": 0.0262,
"step": 1425
},
{
"epoch": 8.68,
"learning_rate": 9e-05,
"loss": 0.0256,
"step": 1450
},
{
"epoch": 8.83,
"learning_rate": 8.973684210526316e-05,
"loss": 0.0286,
"step": 1475
},
{
"epoch": 8.98,
"learning_rate": 8.947368421052632e-05,
"loss": 0.0265,
"step": 1500
},
{
"epoch": 9.13,
"learning_rate": 8.921052631578948e-05,
"loss": 0.0208,
"step": 1525
},
{
"epoch": 9.28,
"learning_rate": 8.894736842105263e-05,
"loss": 0.0237,
"step": 1550
},
{
"epoch": 9.43,
"learning_rate": 8.868421052631579e-05,
"loss": 0.0235,
"step": 1575
},
{
"epoch": 9.58,
"learning_rate": 8.842105263157894e-05,
"loss": 0.0224,
"step": 1600
},
{
"epoch": 9.73,
"learning_rate": 8.81578947368421e-05,
"loss": 0.024,
"step": 1625
},
{
"epoch": 9.88,
"learning_rate": 8.789473684210526e-05,
"loss": 0.0246,
"step": 1650
},
{
"epoch": 10.03,
"learning_rate": 8.763157894736841e-05,
"loss": 0.0248,
"step": 1675
},
{
"epoch": 10.18,
"learning_rate": 8.736842105263158e-05,
"loss": 0.0173,
"step": 1700
},
{
"epoch": 10.33,
"learning_rate": 8.710526315789474e-05,
"loss": 0.0174,
"step": 1725
},
{
"epoch": 10.48,
"learning_rate": 8.68421052631579e-05,
"loss": 0.0215,
"step": 1750
},
{
"epoch": 10.63,
"learning_rate": 8.657894736842105e-05,
"loss": 0.0236,
"step": 1775
},
{
"epoch": 10.78,
"learning_rate": 8.631578947368421e-05,
"loss": 0.0203,
"step": 1800
},
{
"epoch": 10.93,
"learning_rate": 8.605263157894738e-05,
"loss": 0.0192,
"step": 1825
},
{
"epoch": 11.08,
"learning_rate": 8.578947368421054e-05,
"loss": 0.0172,
"step": 1850
},
{
"epoch": 11.23,
"learning_rate": 8.552631578947369e-05,
"loss": 0.0197,
"step": 1875
},
{
"epoch": 11.38,
"learning_rate": 8.526315789473685e-05,
"loss": 0.0188,
"step": 1900
},
{
"epoch": 11.53,
"learning_rate": 8.5e-05,
"loss": 0.0197,
"step": 1925
},
{
"epoch": 11.68,
"learning_rate": 8.473684210526316e-05,
"loss": 0.0197,
"step": 1950
},
{
"epoch": 11.83,
"learning_rate": 8.447368421052632e-05,
"loss": 0.0212,
"step": 1975
},
{
"epoch": 11.98,
"learning_rate": 8.421052631578948e-05,
"loss": 0.0183,
"step": 2000
},
{
"epoch": 11.98,
"eval_loss": 0.49957072734832764,
"eval_runtime": 578.898,
"eval_samples_per_second": 0.724,
"eval_steps_per_second": 0.092,
"eval_wer": 38.2982234200539,
"step": 2000
},
{
"epoch": 12.13,
"learning_rate": 8.394736842105263e-05,
"loss": 0.0152,
"step": 2025
},
{
"epoch": 12.28,
"learning_rate": 8.36842105263158e-05,
"loss": 0.0127,
"step": 2050
},
{
"epoch": 12.43,
"learning_rate": 8.342105263157896e-05,
"loss": 0.0162,
"step": 2075
},
{
"epoch": 12.57,
"learning_rate": 8.315789473684212e-05,
"loss": 0.016,
"step": 2100
},
{
"epoch": 12.72,
"learning_rate": 8.289473684210527e-05,
"loss": 0.0146,
"step": 2125
},
{
"epoch": 12.87,
"learning_rate": 8.263157894736843e-05,
"loss": 0.0202,
"step": 2150
},
{
"epoch": 13.02,
"learning_rate": 8.236842105263158e-05,
"loss": 0.0208,
"step": 2175
},
{
"epoch": 13.17,
"learning_rate": 8.210526315789474e-05,
"loss": 0.0218,
"step": 2200
},
{
"epoch": 13.32,
"learning_rate": 8.18421052631579e-05,
"loss": 0.0191,
"step": 2225
},
{
"epoch": 13.47,
"learning_rate": 8.157894736842105e-05,
"loss": 0.0182,
"step": 2250
},
{
"epoch": 13.62,
"learning_rate": 8.131578947368421e-05,
"loss": 0.02,
"step": 2275
},
{
"epoch": 13.77,
"learning_rate": 8.105263157894737e-05,
"loss": 0.0227,
"step": 2300
},
{
"epoch": 13.92,
"learning_rate": 8.078947368421052e-05,
"loss": 0.0215,
"step": 2325
},
{
"epoch": 14.07,
"learning_rate": 8.052631578947368e-05,
"loss": 0.0207,
"step": 2350
},
{
"epoch": 14.22,
"learning_rate": 8.026315789473685e-05,
"loss": 0.0174,
"step": 2375
},
{
"epoch": 14.37,
"learning_rate": 8e-05,
"loss": 0.0163,
"step": 2400
},
{
"epoch": 14.52,
"learning_rate": 7.973684210526316e-05,
"loss": 0.0161,
"step": 2425
},
{
"epoch": 14.67,
"learning_rate": 7.947368421052632e-05,
"loss": 0.014,
"step": 2450
},
{
"epoch": 14.82,
"learning_rate": 7.921052631578948e-05,
"loss": 0.0157,
"step": 2475
},
{
"epoch": 14.97,
"learning_rate": 7.894736842105263e-05,
"loss": 0.0171,
"step": 2500
},
{
"epoch": 15.12,
"learning_rate": 7.868421052631579e-05,
"loss": 0.0111,
"step": 2525
},
{
"epoch": 15.27,
"learning_rate": 7.842105263157895e-05,
"loss": 0.0119,
"step": 2550
},
{
"epoch": 15.42,
"learning_rate": 7.81578947368421e-05,
"loss": 0.0136,
"step": 2575
},
{
"epoch": 15.57,
"learning_rate": 7.789473684210526e-05,
"loss": 0.0148,
"step": 2600
},
{
"epoch": 15.72,
"learning_rate": 7.763157894736843e-05,
"loss": 0.0123,
"step": 2625
},
{
"epoch": 15.87,
"learning_rate": 7.736842105263159e-05,
"loss": 0.0108,
"step": 2650
},
{
"epoch": 16.02,
"learning_rate": 7.710526315789474e-05,
"loss": 0.013,
"step": 2675
},
{
"epoch": 16.17,
"learning_rate": 7.68421052631579e-05,
"loss": 0.0106,
"step": 2700
},
{
"epoch": 16.32,
"learning_rate": 7.657894736842105e-05,
"loss": 0.0123,
"step": 2725
},
{
"epoch": 16.47,
"learning_rate": 7.631578947368422e-05,
"loss": 0.0152,
"step": 2750
},
{
"epoch": 16.62,
"learning_rate": 7.605263157894738e-05,
"loss": 0.0126,
"step": 2775
},
{
"epoch": 16.77,
"learning_rate": 7.578947368421054e-05,
"loss": 0.0137,
"step": 2800
},
{
"epoch": 16.92,
"learning_rate": 7.55263157894737e-05,
"loss": 0.0163,
"step": 2825
},
{
"epoch": 17.07,
"learning_rate": 7.526315789473685e-05,
"loss": 0.0132,
"step": 2850
},
{
"epoch": 17.22,
"learning_rate": 7.500000000000001e-05,
"loss": 0.011,
"step": 2875
},
{
"epoch": 17.37,
"learning_rate": 7.473684210526316e-05,
"loss": 0.0094,
"step": 2900
},
{
"epoch": 17.51,
"learning_rate": 7.447368421052632e-05,
"loss": 0.0094,
"step": 2925
},
{
"epoch": 17.66,
"learning_rate": 7.421052631578948e-05,
"loss": 0.0107,
"step": 2950
},
{
"epoch": 17.81,
"learning_rate": 7.394736842105263e-05,
"loss": 0.011,
"step": 2975
},
{
"epoch": 17.96,
"learning_rate": 7.368421052631579e-05,
"loss": 0.012,
"step": 3000
},
{
"epoch": 17.96,
"eval_loss": 0.5328223705291748,
"eval_runtime": 578.3593,
"eval_samples_per_second": 0.724,
"eval_steps_per_second": 0.092,
"eval_wer": 38.54023431054397,
"step": 3000
},
{
"epoch": 18.11,
"learning_rate": 7.342105263157895e-05,
"loss": 0.0094,
"step": 3025
},
{
"epoch": 18.26,
"learning_rate": 7.315789473684212e-05,
"loss": 0.0085,
"step": 3050
},
{
"epoch": 18.41,
"learning_rate": 7.289473684210527e-05,
"loss": 0.0089,
"step": 3075
},
{
"epoch": 18.56,
"learning_rate": 7.263157894736843e-05,
"loss": 0.0101,
"step": 3100
},
{
"epoch": 18.71,
"learning_rate": 7.236842105263159e-05,
"loss": 0.0091,
"step": 3125
},
{
"epoch": 18.86,
"learning_rate": 7.210526315789474e-05,
"loss": 0.0106,
"step": 3150
},
{
"epoch": 19.01,
"learning_rate": 7.18421052631579e-05,
"loss": 0.01,
"step": 3175
},
{
"epoch": 19.16,
"learning_rate": 7.157894736842105e-05,
"loss": 0.0067,
"step": 3200
},
{
"epoch": 19.31,
"learning_rate": 7.131578947368421e-05,
"loss": 0.0077,
"step": 3225
},
{
"epoch": 19.46,
"learning_rate": 7.105263157894737e-05,
"loss": 0.0072,
"step": 3250
},
{
"epoch": 19.61,
"learning_rate": 7.078947368421052e-05,
"loss": 0.0066,
"step": 3275
},
{
"epoch": 19.76,
"learning_rate": 7.052631578947368e-05,
"loss": 0.0073,
"step": 3300
},
{
"epoch": 19.91,
"learning_rate": 7.026315789473684e-05,
"loss": 0.0079,
"step": 3325
},
{
"epoch": 20.06,
"learning_rate": 7e-05,
"loss": 0.0101,
"step": 3350
},
{
"epoch": 20.21,
"learning_rate": 6.973684210526315e-05,
"loss": 0.0075,
"step": 3375
},
{
"epoch": 20.36,
"learning_rate": 6.947368421052632e-05,
"loss": 0.0096,
"step": 3400
},
{
"epoch": 20.51,
"learning_rate": 6.921052631578948e-05,
"loss": 0.0095,
"step": 3425
},
{
"epoch": 20.66,
"learning_rate": 6.894736842105263e-05,
"loss": 0.0105,
"step": 3450
},
{
"epoch": 20.81,
"learning_rate": 6.868421052631579e-05,
"loss": 0.0112,
"step": 3475
},
{
"epoch": 20.96,
"learning_rate": 6.842105263157895e-05,
"loss": 0.0116,
"step": 3500
},
{
"epoch": 21.11,
"learning_rate": 6.81578947368421e-05,
"loss": 0.0088,
"step": 3525
},
{
"epoch": 21.26,
"learning_rate": 6.789473684210527e-05,
"loss": 0.01,
"step": 3550
},
{
"epoch": 21.41,
"learning_rate": 6.763157894736843e-05,
"loss": 0.01,
"step": 3575
},
{
"epoch": 21.56,
"learning_rate": 6.736842105263159e-05,
"loss": 0.0083,
"step": 3600
},
{
"epoch": 21.71,
"learning_rate": 6.710526315789474e-05,
"loss": 0.0088,
"step": 3625
},
{
"epoch": 21.86,
"learning_rate": 6.68421052631579e-05,
"loss": 0.0086,
"step": 3650
},
{
"epoch": 22.01,
"learning_rate": 6.657894736842106e-05,
"loss": 0.0078,
"step": 3675
},
{
"epoch": 22.16,
"learning_rate": 6.631578947368421e-05,
"loss": 0.0061,
"step": 3700
},
{
"epoch": 22.31,
"learning_rate": 6.605263157894738e-05,
"loss": 0.0073,
"step": 3725
},
{
"epoch": 22.46,
"learning_rate": 6.578947368421054e-05,
"loss": 0.0071,
"step": 3750
},
{
"epoch": 22.6,
"learning_rate": 6.55263157894737e-05,
"loss": 0.0085,
"step": 3775
},
{
"epoch": 22.75,
"learning_rate": 6.526315789473685e-05,
"loss": 0.0082,
"step": 3800
},
{
"epoch": 22.9,
"learning_rate": 6.500000000000001e-05,
"loss": 0.0098,
"step": 3825
},
{
"epoch": 23.05,
"learning_rate": 6.473684210526316e-05,
"loss": 0.0087,
"step": 3850
},
{
"epoch": 23.2,
"learning_rate": 6.447368421052632e-05,
"loss": 0.0087,
"step": 3875
},
{
"epoch": 23.35,
"learning_rate": 6.421052631578948e-05,
"loss": 0.0076,
"step": 3900
},
{
"epoch": 23.5,
"learning_rate": 6.394736842105263e-05,
"loss": 0.0077,
"step": 3925
},
{
"epoch": 23.65,
"learning_rate": 6.368421052631579e-05,
"loss": 0.0081,
"step": 3950
},
{
"epoch": 23.8,
"learning_rate": 6.342105263157895e-05,
"loss": 0.0083,
"step": 3975
},
{
"epoch": 23.95,
"learning_rate": 6.31578947368421e-05,
"loss": 0.0091,
"step": 4000
},
{
"epoch": 23.95,
"eval_loss": 0.5618667602539062,
"eval_runtime": 582.8261,
"eval_samples_per_second": 0.719,
"eval_steps_per_second": 0.091,
"eval_wer": 38.127715747208626,
"step": 4000
}
],
"logging_steps": 25,
"max_steps": 10000,
"num_input_tokens_seen": 0,
"num_train_epochs": 60,
"save_steps": 1000,
"total_flos": 2.1720546423668736e+20,
"train_batch_size": 16,
"trial_name": null,
"trial_params": null
}