aradia-ctc-v1 / trainer_state.json
Abdulwahab Sahyoun
End of training
3637072
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 14.781758957654723,
"global_step": 6800,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.22,
"eval_loss": 5.188871383666992,
"eval_runtime": 204.5528,
"eval_samples_per_second": 23.95,
"eval_steps_per_second": 1.501,
"eval_wer": 1.0,
"step": 100
},
{
"epoch": 0.43,
"eval_loss": 3.1129117012023926,
"eval_runtime": 202.5006,
"eval_samples_per_second": 24.193,
"eval_steps_per_second": 1.516,
"eval_wer": 1.0,
"step": 200
},
{
"epoch": 0.65,
"eval_loss": 3.050328254699707,
"eval_runtime": 212.2496,
"eval_samples_per_second": 23.081,
"eval_steps_per_second": 1.446,
"eval_wer": 1.0,
"step": 300
},
{
"epoch": 0.87,
"eval_loss": 3.027885913848877,
"eval_runtime": 207.467,
"eval_samples_per_second": 23.613,
"eval_steps_per_second": 1.48,
"eval_wer": 1.0,
"step": 400
},
{
"epoch": 1.09,
"learning_rate": 0.00029759999999999997,
"loss": 6.2756,
"step": 500
},
{
"epoch": 1.09,
"eval_loss": 2.996450662612915,
"eval_runtime": 204.103,
"eval_samples_per_second": 24.003,
"eval_steps_per_second": 1.504,
"eval_wer": 1.0,
"step": 500
},
{
"epoch": 1.3,
"eval_loss": 2.3617873191833496,
"eval_runtime": 203.0453,
"eval_samples_per_second": 24.128,
"eval_steps_per_second": 1.512,
"eval_wer": 0.999270151916527,
"step": 600
},
{
"epoch": 1.52,
"eval_loss": 1.2715116739273071,
"eval_runtime": 201.791,
"eval_samples_per_second": 24.278,
"eval_steps_per_second": 1.521,
"eval_wer": 0.8758177001675947,
"step": 700
},
{
"epoch": 1.74,
"eval_loss": 0.9970846772193909,
"eval_runtime": 204.0755,
"eval_samples_per_second": 24.006,
"eval_steps_per_second": 1.504,
"eval_wer": 0.7156295615505217,
"step": 800
},
{
"epoch": 1.96,
"eval_loss": 0.8927041292190552,
"eval_runtime": 204.7029,
"eval_samples_per_second": 23.932,
"eval_steps_per_second": 1.5,
"eval_wer": 0.6382116018813861,
"step": 900
},
{
"epoch": 2.17,
"learning_rate": 0.00027675,
"loss": 1.712,
"step": 1000
},
{
"epoch": 2.17,
"eval_loss": 0.8251833319664001,
"eval_runtime": 201.65,
"eval_samples_per_second": 24.295,
"eval_steps_per_second": 1.522,
"eval_wer": 0.5926366437800724,
"step": 1000
},
{
"epoch": 2.39,
"eval_loss": 0.7793710827827454,
"eval_runtime": 203.5019,
"eval_samples_per_second": 24.073,
"eval_steps_per_second": 1.509,
"eval_wer": 0.5433854138508948,
"step": 1100
},
{
"epoch": 2.61,
"eval_loss": 0.7556658983230591,
"eval_runtime": 204.1458,
"eval_samples_per_second": 23.998,
"eval_steps_per_second": 1.504,
"eval_wer": 0.509244742390658,
"step": 1200
},
{
"epoch": 2.83,
"eval_loss": 0.7346892952919006,
"eval_runtime": 201.906,
"eval_samples_per_second": 24.264,
"eval_steps_per_second": 1.521,
"eval_wer": 0.5203005892847489,
"step": 1300
},
{
"epoch": 3.04,
"eval_loss": 0.7189395427703857,
"eval_runtime": 202.2047,
"eval_samples_per_second": 24.228,
"eval_steps_per_second": 1.518,
"eval_wer": 0.49286370762826404,
"step": 1400
},
{
"epoch": 3.26,
"learning_rate": 0.00025331249999999996,
"loss": 0.9305,
"step": 1500
},
{
"epoch": 3.26,
"eval_loss": 0.6819509267807007,
"eval_runtime": 201.6073,
"eval_samples_per_second": 24.3,
"eval_steps_per_second": 1.523,
"eval_wer": 0.45947991566199925,
"step": 1500
},
{
"epoch": 3.48,
"eval_loss": 0.679171085357666,
"eval_runtime": 201.2709,
"eval_samples_per_second": 24.34,
"eval_steps_per_second": 1.525,
"eval_wer": 0.4504243931448343,
"step": 1600
},
{
"epoch": 3.69,
"eval_loss": 0.6596325635910034,
"eval_runtime": 201.0715,
"eval_samples_per_second": 24.364,
"eval_steps_per_second": 1.527,
"eval_wer": 0.44423420014056336,
"step": 1700
},
{
"epoch": 3.91,
"eval_loss": 0.6755712628364563,
"eval_runtime": 202.0726,
"eval_samples_per_second": 24.244,
"eval_steps_per_second": 1.519,
"eval_wer": 0.4431529437206033,
"step": 1800
},
{
"epoch": 4.13,
"eval_loss": 0.6663188338279724,
"eval_runtime": 200.5093,
"eval_samples_per_second": 24.433,
"eval_steps_per_second": 1.531,
"eval_wer": 0.43920635778774936,
"step": 1900
},
{
"epoch": 4.35,
"learning_rate": 0.00022987499999999996,
"loss": 0.737,
"step": 2000
},
{
"epoch": 4.35,
"eval_loss": 0.6478506326675415,
"eval_runtime": 202.9556,
"eval_samples_per_second": 24.138,
"eval_steps_per_second": 1.513,
"eval_wer": 0.43720603341082337,
"step": 2000
},
{
"epoch": 4.56,
"eval_loss": 0.6353247761726379,
"eval_runtime": 202.5451,
"eval_samples_per_second": 24.187,
"eval_steps_per_second": 1.516,
"eval_wer": 0.4203384332594475,
"step": 2100
},
{
"epoch": 4.78,
"eval_loss": 0.6251159906387329,
"eval_runtime": 204.4658,
"eval_samples_per_second": 23.96,
"eval_steps_per_second": 1.501,
"eval_wer": 0.4087960209763746,
"step": 2200
},
{
"epoch": 5.0,
"eval_loss": 0.6208750605583191,
"eval_runtime": 204.1125,
"eval_samples_per_second": 24.001,
"eval_steps_per_second": 1.504,
"eval_wer": 0.4176893550305455,
"step": 2300
},
{
"epoch": 5.22,
"eval_loss": 0.6639252305030823,
"eval_runtime": 235.6403,
"eval_samples_per_second": 20.79,
"eval_steps_per_second": 1.303,
"eval_wer": 0.40941774341785153,
"step": 2400
},
{
"epoch": 5.43,
"learning_rate": 0.00020643749999999997,
"loss": 0.6247,
"step": 2500
},
{
"epoch": 5.43,
"eval_loss": 0.6407893300056458,
"eval_runtime": 278.1573,
"eval_samples_per_second": 17.612,
"eval_steps_per_second": 1.104,
"eval_wer": 0.39701032599881064,
"step": 2500
},
{
"epoch": 5.65,
"eval_loss": 0.6373021006584167,
"eval_runtime": 207.7541,
"eval_samples_per_second": 23.581,
"eval_steps_per_second": 1.478,
"eval_wer": 0.39322592852895066,
"step": 2600
},
{
"epoch": 5.87,
"eval_loss": 0.6411147713661194,
"eval_runtime": 228.4908,
"eval_samples_per_second": 21.441,
"eval_steps_per_second": 1.344,
"eval_wer": 0.39276639455046763,
"step": 2700
},
{
"epoch": 6.09,
"eval_loss": 0.6378452181816101,
"eval_runtime": 202.0494,
"eval_samples_per_second": 24.247,
"eval_steps_per_second": 1.519,
"eval_wer": 0.38971184516408064,
"step": 2800
},
{
"epoch": 6.3,
"eval_loss": 0.6396178007125854,
"eval_runtime": 201.6642,
"eval_samples_per_second": 24.293,
"eval_steps_per_second": 1.522,
"eval_wer": 0.39287452019246366,
"step": 2900
},
{
"epoch": 6.52,
"learning_rate": 0.00018299999999999998,
"loss": 0.5443,
"step": 3000
},
{
"epoch": 6.52,
"eval_loss": 0.6544364094734192,
"eval_runtime": 201.2122,
"eval_samples_per_second": 24.347,
"eval_steps_per_second": 1.526,
"eval_wer": 0.38641401308320267,
"step": 3000
},
{
"epoch": 6.74,
"eval_loss": 0.6217912435531616,
"eval_runtime": 202.6291,
"eval_samples_per_second": 24.177,
"eval_steps_per_second": 1.515,
"eval_wer": 0.37857490403849275,
"step": 3100
},
{
"epoch": 6.96,
"eval_loss": 0.6200193166732788,
"eval_runtime": 202.7362,
"eval_samples_per_second": 24.164,
"eval_steps_per_second": 1.514,
"eval_wer": 0.37843974698599775,
"step": 3200
},
{
"epoch": 7.17,
"eval_loss": 0.6156935691833496,
"eval_runtime": 203.722,
"eval_samples_per_second": 24.047,
"eval_steps_per_second": 1.507,
"eval_wer": 0.3790614694274747,
"step": 3300
},
{
"epoch": 7.39,
"eval_loss": 0.6316530704498291,
"eval_runtime": 202.1843,
"eval_samples_per_second": 24.23,
"eval_steps_per_second": 1.518,
"eval_wer": 0.37979131751094775,
"step": 3400
},
{
"epoch": 7.61,
"learning_rate": 0.00015956249999999998,
"loss": 0.4845,
"step": 3500
},
{
"epoch": 7.61,
"eval_loss": 0.6540383100509644,
"eval_runtime": 203.1885,
"eval_samples_per_second": 24.111,
"eval_steps_per_second": 1.511,
"eval_wer": 0.37711520787154673,
"step": 3500
},
{
"epoch": 7.83,
"eval_loss": 0.6435653567314148,
"eval_runtime": 201.9898,
"eval_samples_per_second": 24.254,
"eval_steps_per_second": 1.52,
"eval_wer": 0.3669784289344218,
"step": 3600
},
{
"epoch": 8.04,
"eval_loss": 0.63348388671875,
"eval_runtime": 201.5967,
"eval_samples_per_second": 24.301,
"eval_steps_per_second": 1.523,
"eval_wer": 0.3694653187003298,
"step": 3700
},
{
"epoch": 8.26,
"eval_loss": 0.657917320728302,
"eval_runtime": 208.9718,
"eval_samples_per_second": 23.443,
"eval_steps_per_second": 1.469,
"eval_wer": 0.36100448721414286,
"step": 3800
},
{
"epoch": 8.48,
"eval_loss": 0.6169739961624146,
"eval_runtime": 203.6316,
"eval_samples_per_second": 24.058,
"eval_steps_per_second": 1.508,
"eval_wer": 0.3613288641401308,
"step": 3900
},
{
"epoch": 8.69,
"learning_rate": 0.000136125,
"loss": 0.4279,
"step": 4000
},
{
"epoch": 8.69,
"eval_loss": 0.6523476243019104,
"eval_runtime": 201.8094,
"eval_samples_per_second": 24.275,
"eval_steps_per_second": 1.521,
"eval_wer": 0.36165324106611885,
"step": 4000
},
{
"epoch": 8.91,
"eval_loss": 0.634904146194458,
"eval_runtime": 202.5275,
"eval_samples_per_second": 24.189,
"eval_steps_per_second": 1.516,
"eval_wer": 0.35773368654376386,
"step": 4100
},
{
"epoch": 9.13,
"eval_loss": 0.6343770027160645,
"eval_runtime": 200.9349,
"eval_samples_per_second": 24.381,
"eval_steps_per_second": 1.528,
"eval_wer": 0.3673298372709088,
"step": 4200
},
{
"epoch": 9.35,
"eval_loss": 0.6215314269065857,
"eval_runtime": 200.3697,
"eval_samples_per_second": 24.45,
"eval_steps_per_second": 1.532,
"eval_wer": 0.3641130994215278,
"step": 4300
},
{
"epoch": 9.56,
"eval_loss": 0.6512671709060669,
"eval_runtime": 201.0806,
"eval_samples_per_second": 24.363,
"eval_steps_per_second": 1.527,
"eval_wer": 0.36081526734064984,
"step": 4400
},
{
"epoch": 9.78,
"learning_rate": 0.00011268749999999998,
"loss": 0.3825,
"step": 4500
},
{
"epoch": 9.78,
"eval_loss": 0.638552725315094,
"eval_runtime": 201.1524,
"eval_samples_per_second": 24.355,
"eval_steps_per_second": 1.526,
"eval_wer": 0.36051792182516085,
"step": 4500
},
{
"epoch": 10.0,
"eval_loss": 0.6724193692207336,
"eval_runtime": 201.6562,
"eval_samples_per_second": 24.294,
"eval_steps_per_second": 1.522,
"eval_wer": 0.3549494512623669,
"step": 4600
},
{
"epoch": 10.22,
"eval_loss": 0.6775844693183899,
"eval_runtime": 201.0063,
"eval_samples_per_second": 24.372,
"eval_steps_per_second": 1.527,
"eval_wer": 0.36016651348867385,
"step": 4700
},
{
"epoch": 10.43,
"eval_loss": 0.673897385597229,
"eval_runtime": 200.6197,
"eval_samples_per_second": 24.419,
"eval_steps_per_second": 1.53,
"eval_wer": 0.3543817916418879,
"step": 4800
},
{
"epoch": 10.65,
"eval_loss": 0.6688181161880493,
"eval_runtime": 202.0195,
"eval_samples_per_second": 24.25,
"eval_steps_per_second": 1.52,
"eval_wer": 0.3557333621668379,
"step": 4900
},
{
"epoch": 10.87,
"learning_rate": 8.924999999999999e-05,
"loss": 0.3477,
"step": 5000
},
{
"epoch": 10.87,
"eval_loss": 0.6674283146858215,
"eval_runtime": 204.4147,
"eval_samples_per_second": 23.966,
"eval_steps_per_second": 1.502,
"eval_wer": 0.3564361788398119,
"step": 5000
},
{
"epoch": 11.09,
"eval_loss": 0.6786395907402039,
"eval_runtime": 201.526,
"eval_samples_per_second": 24.31,
"eval_steps_per_second": 1.523,
"eval_wer": 0.3475969076066389,
"step": 5100
},
{
"epoch": 11.3,
"eval_loss": 0.6817718148231506,
"eval_runtime": 201.6908,
"eval_samples_per_second": 24.29,
"eval_steps_per_second": 1.522,
"eval_wer": 0.34778612748013193,
"step": 5200
},
{
"epoch": 11.52,
"eval_loss": 0.6873587965965271,
"eval_runtime": 199.7159,
"eval_samples_per_second": 24.53,
"eval_steps_per_second": 1.537,
"eval_wer": 0.3469751851651619,
"step": 5300
},
{
"epoch": 11.74,
"eval_loss": 0.6992678642272949,
"eval_runtime": 201.0468,
"eval_samples_per_second": 24.367,
"eval_steps_per_second": 1.527,
"eval_wer": 0.3424068767908309,
"step": 5400
},
{
"epoch": 11.96,
"learning_rate": 6.58125e-05,
"loss": 0.3101,
"step": 5500
},
{
"epoch": 11.96,
"eval_loss": 0.6949637532234192,
"eval_runtime": 201.2671,
"eval_samples_per_second": 24.341,
"eval_steps_per_second": 1.525,
"eval_wer": 0.34040655241390494,
"step": 5500
},
{
"epoch": 12.17,
"eval_loss": 0.6872305274009705,
"eval_runtime": 205.5246,
"eval_samples_per_second": 23.837,
"eval_steps_per_second": 1.494,
"eval_wer": 0.34062280369789694,
"step": 5600
},
{
"epoch": 12.39,
"eval_loss": 0.6846413016319275,
"eval_runtime": 201.5806,
"eval_samples_per_second": 24.303,
"eval_steps_per_second": 1.523,
"eval_wer": 0.34237984538033195,
"step": 5700
},
{
"epoch": 12.61,
"eval_loss": 0.7050597667694092,
"eval_runtime": 201.3274,
"eval_samples_per_second": 24.333,
"eval_steps_per_second": 1.525,
"eval_wer": 0.34046061523490295,
"step": 5800
},
{
"epoch": 12.83,
"eval_loss": 0.7051235437393188,
"eval_runtime": 201.5539,
"eval_samples_per_second": 24.306,
"eval_steps_per_second": 1.523,
"eval_wer": 0.337811537006001,
"step": 5900
},
{
"epoch": 13.04,
"learning_rate": 4.237499999999999e-05,
"loss": 0.2859,
"step": 6000
},
{
"epoch": 13.04,
"eval_loss": 0.6955370306968689,
"eval_runtime": 201.221,
"eval_samples_per_second": 24.346,
"eval_steps_per_second": 1.526,
"eval_wer": 0.34032545818240795,
"step": 6000
},
{
"epoch": 13.26,
"eval_loss": 0.7115095257759094,
"eval_runtime": 201.4349,
"eval_samples_per_second": 24.321,
"eval_steps_per_second": 1.524,
"eval_wer": 0.33902795047845596,
"step": 6100
},
{
"epoch": 13.48,
"eval_loss": 0.7074177861213684,
"eval_runtime": 200.2975,
"eval_samples_per_second": 24.459,
"eval_steps_per_second": 1.533,
"eval_wer": 0.338433259447478,
"step": 6200
},
{
"epoch": 13.69,
"eval_loss": 0.7001507878303528,
"eval_runtime": 201.4343,
"eval_samples_per_second": 24.321,
"eval_steps_per_second": 1.524,
"eval_wer": 0.33756825431151,
"step": 6300
},
{
"epoch": 13.91,
"eval_loss": 0.7170686721801758,
"eval_runtime": 203.7443,
"eval_samples_per_second": 24.045,
"eval_steps_per_second": 1.507,
"eval_wer": 0.336027463913067,
"step": 6400
},
{
"epoch": 14.13,
"learning_rate": 1.89375e-05,
"loss": 0.2714,
"step": 6500
},
{
"epoch": 14.13,
"eval_loss": 0.7192927002906799,
"eval_runtime": 200.9828,
"eval_samples_per_second": 24.375,
"eval_steps_per_second": 1.527,
"eval_wer": 0.33405417094663997,
"step": 6500
},
{
"epoch": 14.35,
"eval_loss": 0.7131814360618591,
"eval_runtime": 199.8266,
"eval_samples_per_second": 24.516,
"eval_steps_per_second": 1.536,
"eval_wer": 0.334729956209115,
"step": 6600
},
{
"epoch": 14.56,
"eval_loss": 0.7184281349182129,
"eval_runtime": 202.0064,
"eval_samples_per_second": 24.252,
"eval_steps_per_second": 1.52,
"eval_wer": 0.335297615829594,
"step": 6700
},
{
"epoch": 14.78,
"eval_loss": 0.7170840501785278,
"eval_runtime": 201.3006,
"eval_samples_per_second": 24.337,
"eval_steps_per_second": 1.525,
"eval_wer": 0.333054008758177,
"step": 6800
},
{
"epoch": 14.78,
"step": 6800,
"total_flos": 2.9460768962998075e+19,
"train_loss": 0.0,
"train_runtime": 70.228,
"train_samples_per_second": 4194.91,
"train_steps_per_second": 65.501
}
],
"max_steps": 4600,
"num_train_epochs": 20,
"total_flos": 2.9460768962998075e+19,
"trial_name": null,
"trial_params": null
}