the-final-whisper / checkpoint-1750 /trainer_state.json
AsemBadr's picture
Training in progress, step 1750, checkpoint
5f6a9b1 verified
raw
history blame
15.9 kB
{
"best_metric": 14.024702866464695,
"best_model_checkpoint": "./the-final-whisper/checkpoint-1750",
"epoch": 2.883031301482702,
"eval_steps": 125,
"global_step": 1750,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.04118616144975288,
"grad_norm": 52.4793701171875,
"learning_rate": 4.2000000000000006e-07,
"loss": 2.7503,
"step": 25
},
{
"epoch": 0.08237232289950576,
"grad_norm": 16.525333404541016,
"learning_rate": 9.200000000000001e-07,
"loss": 2.0994,
"step": 50
},
{
"epoch": 0.12355848434925865,
"grad_norm": 14.271560668945312,
"learning_rate": 1.42e-06,
"loss": 1.3213,
"step": 75
},
{
"epoch": 0.16474464579901152,
"grad_norm": 8.628070831298828,
"learning_rate": 1.9200000000000003e-06,
"loss": 0.7342,
"step": 100
},
{
"epoch": 0.20593080724876442,
"grad_norm": 8.307537078857422,
"learning_rate": 2.42e-06,
"loss": 0.5674,
"step": 125
},
{
"epoch": 0.20593080724876442,
"eval_loss": 0.6090311408042908,
"eval_runtime": 2805.8489,
"eval_samples_per_second": 1.442,
"eval_steps_per_second": 0.18,
"eval_wer": 45.15497553017945,
"step": 125
},
{
"epoch": 0.2471169686985173,
"grad_norm": 6.402714729309082,
"learning_rate": 2.92e-06,
"loss": 0.4853,
"step": 150
},
{
"epoch": 0.2883031301482702,
"grad_norm": 7.138896942138672,
"learning_rate": 3.4200000000000007e-06,
"loss": 0.424,
"step": 175
},
{
"epoch": 0.32948929159802304,
"grad_norm": 6.78637170791626,
"learning_rate": 3.920000000000001e-06,
"loss": 0.3443,
"step": 200
},
{
"epoch": 0.37067545304777594,
"grad_norm": 6.042959690093994,
"learning_rate": 4.42e-06,
"loss": 0.2769,
"step": 225
},
{
"epoch": 0.41186161449752884,
"grad_norm": 2.8671882152557373,
"learning_rate": 4.92e-06,
"loss": 0.1545,
"step": 250
},
{
"epoch": 0.41186161449752884,
"eval_loss": 0.18588024377822876,
"eval_runtime": 2810.3027,
"eval_samples_per_second": 1.439,
"eval_steps_per_second": 0.18,
"eval_wer": 30.12351433232347,
"step": 250
},
{
"epoch": 0.45304777594728174,
"grad_norm": 4.210362911224365,
"learning_rate": 5.420000000000001e-06,
"loss": 0.094,
"step": 275
},
{
"epoch": 0.4942339373970346,
"grad_norm": 3.7346065044403076,
"learning_rate": 5.92e-06,
"loss": 0.0858,
"step": 300
},
{
"epoch": 0.5354200988467874,
"grad_norm": 2.7393386363983154,
"learning_rate": 6.42e-06,
"loss": 0.0745,
"step": 325
},
{
"epoch": 0.5766062602965404,
"grad_norm": 2.8941423892974854,
"learning_rate": 6.92e-06,
"loss": 0.0631,
"step": 350
},
{
"epoch": 0.6177924217462932,
"grad_norm": 3.1176912784576416,
"learning_rate": 7.420000000000001e-06,
"loss": 0.0658,
"step": 375
},
{
"epoch": 0.6177924217462932,
"eval_loss": 0.1334521770477295,
"eval_runtime": 2821.0386,
"eval_samples_per_second": 1.434,
"eval_steps_per_second": 0.179,
"eval_wer": 22.91773479375437,
"step": 375
},
{
"epoch": 0.6589785831960461,
"grad_norm": 3.2934091091156006,
"learning_rate": 7.92e-06,
"loss": 0.0552,
"step": 400
},
{
"epoch": 0.700164744645799,
"grad_norm": 3.6386263370513916,
"learning_rate": 8.42e-06,
"loss": 0.0529,
"step": 425
},
{
"epoch": 0.7413509060955519,
"grad_norm": 1.1384811401367188,
"learning_rate": 8.920000000000001e-06,
"loss": 0.0418,
"step": 450
},
{
"epoch": 0.7825370675453048,
"grad_norm": 4.086984634399414,
"learning_rate": 9.42e-06,
"loss": 0.0493,
"step": 475
},
{
"epoch": 0.8237232289950577,
"grad_norm": 2.331613540649414,
"learning_rate": 9.920000000000002e-06,
"loss": 0.0378,
"step": 500
},
{
"epoch": 0.8237232289950577,
"eval_loss": 0.12090734392404556,
"eval_runtime": 2820.3109,
"eval_samples_per_second": 1.434,
"eval_steps_per_second": 0.179,
"eval_wer": 20.77837333954789,
"step": 500
},
{
"epoch": 0.8649093904448105,
"grad_norm": 2.7532379627227783,
"learning_rate": 9.58e-06,
"loss": 0.0365,
"step": 525
},
{
"epoch": 0.9060955518945635,
"grad_norm": 3.022937774658203,
"learning_rate": 9.080000000000001e-06,
"loss": 0.0414,
"step": 550
},
{
"epoch": 0.9472817133443163,
"grad_norm": 2.641817092895508,
"learning_rate": 8.580000000000001e-06,
"loss": 0.0319,
"step": 575
},
{
"epoch": 0.9884678747940692,
"grad_norm": 3.3846614360809326,
"learning_rate": 8.08e-06,
"loss": 0.0272,
"step": 600
},
{
"epoch": 1.029654036243822,
"grad_norm": 4.221302032470703,
"learning_rate": 7.58e-06,
"loss": 0.022,
"step": 625
},
{
"epoch": 1.029654036243822,
"eval_loss": 0.10999257117509842,
"eval_runtime": 2822.9012,
"eval_samples_per_second": 1.433,
"eval_steps_per_second": 0.179,
"eval_wer": 18.08436261943603,
"step": 625
},
{
"epoch": 1.0708401976935749,
"grad_norm": 4.843472480773926,
"learning_rate": 7.08e-06,
"loss": 0.021,
"step": 650
},
{
"epoch": 1.112026359143328,
"grad_norm": 1.8119359016418457,
"learning_rate": 6.5800000000000005e-06,
"loss": 0.0141,
"step": 675
},
{
"epoch": 1.1532125205930808,
"grad_norm": 1.112451195716858,
"learning_rate": 6.08e-06,
"loss": 0.0148,
"step": 700
},
{
"epoch": 1.1943986820428336,
"grad_norm": 0.6692703366279602,
"learning_rate": 5.580000000000001e-06,
"loss": 0.0126,
"step": 725
},
{
"epoch": 1.2355848434925865,
"grad_norm": 1.457305908203125,
"learning_rate": 5.0800000000000005e-06,
"loss": 0.0097,
"step": 750
},
{
"epoch": 1.2355848434925865,
"eval_loss": 0.09214470535516739,
"eval_runtime": 2819.3708,
"eval_samples_per_second": 1.435,
"eval_steps_per_second": 0.179,
"eval_wer": 15.441621999533908,
"step": 750
},
{
"epoch": 1.2767710049423393,
"grad_norm": 2.9795682430267334,
"learning_rate": 4.58e-06,
"loss": 0.0105,
"step": 775
},
{
"epoch": 1.3179571663920924,
"grad_norm": 1.0585089921951294,
"learning_rate": 4.08e-06,
"loss": 0.0083,
"step": 800
},
{
"epoch": 1.3591433278418452,
"grad_norm": 2.7793498039245605,
"learning_rate": 3.58e-06,
"loss": 0.0113,
"step": 825
},
{
"epoch": 1.400329489291598,
"grad_norm": 0.24739088118076324,
"learning_rate": 3.08e-06,
"loss": 0.0106,
"step": 850
},
{
"epoch": 1.441515650741351,
"grad_norm": 0.5349621176719666,
"learning_rate": 2.5800000000000003e-06,
"loss": 0.0078,
"step": 875
},
{
"epoch": 1.441515650741351,
"eval_loss": 0.08723945915699005,
"eval_runtime": 2823.033,
"eval_samples_per_second": 1.433,
"eval_steps_per_second": 0.179,
"eval_wer": 15.530179445350733,
"step": 875
},
{
"epoch": 1.4827018121911038,
"grad_norm": 3.1261961460113525,
"learning_rate": 2.08e-06,
"loss": 0.0098,
"step": 900
},
{
"epoch": 1.5238879736408566,
"grad_norm": 0.6137211322784424,
"learning_rate": 1.5800000000000001e-06,
"loss": 0.0092,
"step": 925
},
{
"epoch": 1.5650741350906094,
"grad_norm": 1.3198140859603882,
"learning_rate": 1.08e-06,
"loss": 0.0088,
"step": 950
},
{
"epoch": 1.6062602965403623,
"grad_norm": 0.7606719136238098,
"learning_rate": 5.800000000000001e-07,
"loss": 0.0132,
"step": 975
},
{
"epoch": 1.6474464579901154,
"grad_norm": 1.69697105884552,
"learning_rate": 8e-08,
"loss": 0.012,
"step": 1000
},
{
"epoch": 1.6474464579901154,
"eval_loss": 0.08294647186994553,
"eval_runtime": 2832.7593,
"eval_samples_per_second": 1.428,
"eval_steps_per_second": 0.179,
"eval_wer": 14.845024469820556,
"step": 1000
},
{
"epoch": 1.6886326194398682,
"grad_norm": 1.8255125284194946,
"learning_rate": 8.842222222222223e-06,
"loss": 0.0097,
"step": 1025
},
{
"epoch": 1.729818780889621,
"grad_norm": 1.6991933584213257,
"learning_rate": 8.786666666666668e-06,
"loss": 0.0128,
"step": 1050
},
{
"epoch": 1.771004942339374,
"grad_norm": 3.427128314971924,
"learning_rate": 8.73111111111111e-06,
"loss": 0.0127,
"step": 1075
},
{
"epoch": 1.812191103789127,
"grad_norm": 1.3395907878875732,
"learning_rate": 8.675555555555556e-06,
"loss": 0.0111,
"step": 1100
},
{
"epoch": 1.8533772652388798,
"grad_norm": 2.087146520614624,
"learning_rate": 8.62e-06,
"loss": 0.0098,
"step": 1125
},
{
"epoch": 1.8533772652388798,
"eval_loss": 0.09520899504423141,
"eval_runtime": 2831.8668,
"eval_samples_per_second": 1.428,
"eval_steps_per_second": 0.179,
"eval_wer": 16.11745513866232,
"step": 1125
},
{
"epoch": 1.8945634266886326,
"grad_norm": 1.362247109413147,
"learning_rate": 8.564444444444445e-06,
"loss": 0.0102,
"step": 1150
},
{
"epoch": 1.9357495881383855,
"grad_norm": 1.9441858530044556,
"learning_rate": 8.50888888888889e-06,
"loss": 0.012,
"step": 1175
},
{
"epoch": 1.9769357495881383,
"grad_norm": 1.1025640964508057,
"learning_rate": 8.453333333333334e-06,
"loss": 0.0094,
"step": 1200
},
{
"epoch": 2.018121911037891,
"grad_norm": 0.26811718940734863,
"learning_rate": 8.397777777777778e-06,
"loss": 0.0103,
"step": 1225
},
{
"epoch": 2.059308072487644,
"grad_norm": 0.3004502058029175,
"learning_rate": 8.342222222222222e-06,
"loss": 0.0038,
"step": 1250
},
{
"epoch": 2.059308072487644,
"eval_loss": 0.09276822954416275,
"eval_runtime": 2807.4549,
"eval_samples_per_second": 1.441,
"eval_steps_per_second": 0.18,
"eval_wer": 15.422978326730366,
"step": 1250
},
{
"epoch": 2.100494233937397,
"grad_norm": 1.0379807949066162,
"learning_rate": 8.286666666666668e-06,
"loss": 0.0064,
"step": 1275
},
{
"epoch": 2.1416803953871497,
"grad_norm": 2.824099063873291,
"learning_rate": 8.231111111111112e-06,
"loss": 0.0063,
"step": 1300
},
{
"epoch": 2.182866556836903,
"grad_norm": 0.2786366939544678,
"learning_rate": 8.175555555555556e-06,
"loss": 0.0057,
"step": 1325
},
{
"epoch": 2.224052718286656,
"grad_norm": 1.7999166250228882,
"learning_rate": 8.120000000000002e-06,
"loss": 0.0057,
"step": 1350
},
{
"epoch": 2.2652388797364087,
"grad_norm": 1.3867113590240479,
"learning_rate": 8.064444444444444e-06,
"loss": 0.0085,
"step": 1375
},
{
"epoch": 2.2652388797364087,
"eval_loss": 0.09214825183153152,
"eval_runtime": 2797.2114,
"eval_samples_per_second": 1.446,
"eval_steps_per_second": 0.181,
"eval_wer": 15.665346073176417,
"step": 1375
},
{
"epoch": 2.3064250411861615,
"grad_norm": 0.2658134996891022,
"learning_rate": 8.00888888888889e-06,
"loss": 0.005,
"step": 1400
},
{
"epoch": 2.3476112026359144,
"grad_norm": 1.3457545042037964,
"learning_rate": 7.953333333333334e-06,
"loss": 0.0053,
"step": 1425
},
{
"epoch": 2.3887973640856672,
"grad_norm": 1.7825807332992554,
"learning_rate": 7.897777777777778e-06,
"loss": 0.0044,
"step": 1450
},
{
"epoch": 2.42998352553542,
"grad_norm": 2.2161757946014404,
"learning_rate": 7.842222222222224e-06,
"loss": 0.0056,
"step": 1475
},
{
"epoch": 2.471169686985173,
"grad_norm": 0.23896081745624542,
"learning_rate": 7.786666666666666e-06,
"loss": 0.0045,
"step": 1500
},
{
"epoch": 2.471169686985173,
"eval_loss": 0.09182223677635193,
"eval_runtime": 2787.3566,
"eval_samples_per_second": 1.451,
"eval_steps_per_second": 0.182,
"eval_wer": 14.448846422745282,
"step": 1500
},
{
"epoch": 2.5123558484349258,
"grad_norm": 0.37352365255355835,
"learning_rate": 7.731111111111112e-06,
"loss": 0.0044,
"step": 1525
},
{
"epoch": 2.5535420098846786,
"grad_norm": 1.8053451776504517,
"learning_rate": 7.675555555555556e-06,
"loss": 0.0056,
"step": 1550
},
{
"epoch": 2.594728171334432,
"grad_norm": 1.4376009702682495,
"learning_rate": 7.620000000000001e-06,
"loss": 0.007,
"step": 1575
},
{
"epoch": 2.6359143327841847,
"grad_norm": 0.46193966269493103,
"learning_rate": 7.564444444444446e-06,
"loss": 0.0052,
"step": 1600
},
{
"epoch": 2.6771004942339376,
"grad_norm": 0.06809718906879425,
"learning_rate": 7.50888888888889e-06,
"loss": 0.0031,
"step": 1625
},
{
"epoch": 2.6771004942339376,
"eval_loss": 0.08786112070083618,
"eval_runtime": 2779.1365,
"eval_samples_per_second": 1.455,
"eval_steps_per_second": 0.182,
"eval_wer": 14.048007457469122,
"step": 1625
},
{
"epoch": 2.7182866556836904,
"grad_norm": 0.12468615174293518,
"learning_rate": 7.453333333333334e-06,
"loss": 0.0032,
"step": 1650
},
{
"epoch": 2.7594728171334433,
"grad_norm": 0.25042837858200073,
"learning_rate": 7.3977777777777786e-06,
"loss": 0.0062,
"step": 1675
},
{
"epoch": 2.800658978583196,
"grad_norm": 1.6082072257995605,
"learning_rate": 7.342222222222223e-06,
"loss": 0.0036,
"step": 1700
},
{
"epoch": 2.841845140032949,
"grad_norm": 0.5134331583976746,
"learning_rate": 7.2866666666666675e-06,
"loss": 0.0057,
"step": 1725
},
{
"epoch": 2.883031301482702,
"grad_norm": 0.7861504554748535,
"learning_rate": 7.231111111111112e-06,
"loss": 0.0059,
"step": 1750
},
{
"epoch": 2.883031301482702,
"eval_loss": 0.09202219545841217,
"eval_runtime": 2759.663,
"eval_samples_per_second": 1.466,
"eval_steps_per_second": 0.183,
"eval_wer": 14.024702866464695,
"step": 1750
}
],
"logging_steps": 25,
"max_steps": 5000,
"num_input_tokens_seen": 0,
"num_train_epochs": 9,
"save_steps": 250,
"total_flos": 8.07808252575744e+18,
"train_batch_size": 16,
"trial_name": null,
"trial_params": null
}