Shukv4 / trainer_state.json
sant
Upload model files
1213cbd verified
{
"best_metric": 85.87243015287297,
"best_model_checkpoint": "./Shukv4/checkpoint-500",
"epoch": 5.376344086021505,
"eval_steps": 500,
"global_step": 1000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.13440860215053763,
"grad_norm": 349994.25,
"learning_rate": 3.125e-06,
"loss": 0.5555,
"step": 25
},
{
"epoch": 0.26881720430107525,
"grad_norm": 296199.8125,
"learning_rate": 6.25e-06,
"loss": 0.3776,
"step": 50
},
{
"epoch": 0.4032258064516129,
"grad_norm": 218550.265625,
"learning_rate": 9.375000000000001e-06,
"loss": 0.286,
"step": 75
},
{
"epoch": 0.5376344086021505,
"grad_norm": 241839.0,
"learning_rate": 1.25e-05,
"loss": 0.2629,
"step": 100
},
{
"epoch": 0.6720430107526881,
"grad_norm": 224651.953125,
"learning_rate": 1.2152777777777779e-05,
"loss": 0.2473,
"step": 125
},
{
"epoch": 0.8064516129032258,
"grad_norm": 247373.3125,
"learning_rate": 1.1805555555555555e-05,
"loss": 0.2225,
"step": 150
},
{
"epoch": 0.9408602150537635,
"grad_norm": 232539.5,
"learning_rate": 1.1458333333333333e-05,
"loss": 0.1814,
"step": 175
},
{
"epoch": 1.075268817204301,
"grad_norm": 208122.359375,
"learning_rate": 1.1111111111111112e-05,
"loss": 0.1379,
"step": 200
},
{
"epoch": 1.2096774193548387,
"grad_norm": 203154.734375,
"learning_rate": 1.076388888888889e-05,
"loss": 0.0926,
"step": 225
},
{
"epoch": 1.3440860215053765,
"grad_norm": 193545.453125,
"learning_rate": 1.0416666666666668e-05,
"loss": 0.0877,
"step": 250
},
{
"epoch": 1.478494623655914,
"grad_norm": 159916.4375,
"learning_rate": 1.0069444444444445e-05,
"loss": 0.0759,
"step": 275
},
{
"epoch": 1.6129032258064515,
"grad_norm": 203974.796875,
"learning_rate": 9.722222222222223e-06,
"loss": 0.0886,
"step": 300
},
{
"epoch": 1.7473118279569892,
"grad_norm": 176038.296875,
"learning_rate": 9.375000000000001e-06,
"loss": 0.0809,
"step": 325
},
{
"epoch": 1.881720430107527,
"grad_norm": 156732.640625,
"learning_rate": 9.027777777777777e-06,
"loss": 0.0598,
"step": 350
},
{
"epoch": 2.0161290322580645,
"grad_norm": 116201.7734375,
"learning_rate": 8.680555555555556e-06,
"loss": 0.0546,
"step": 375
},
{
"epoch": 2.150537634408602,
"grad_norm": 81153.625,
"learning_rate": 8.333333333333334e-06,
"loss": 0.0281,
"step": 400
},
{
"epoch": 2.28494623655914,
"grad_norm": 95937.9765625,
"learning_rate": 7.98611111111111e-06,
"loss": 0.0328,
"step": 425
},
{
"epoch": 2.4193548387096775,
"grad_norm": 105021.359375,
"learning_rate": 7.63888888888889e-06,
"loss": 0.0271,
"step": 450
},
{
"epoch": 2.553763440860215,
"grad_norm": 73798.6796875,
"learning_rate": 7.2916666666666674e-06,
"loss": 0.0317,
"step": 475
},
{
"epoch": 2.688172043010753,
"grad_norm": 111273.5546875,
"learning_rate": 6.944444444444445e-06,
"loss": 0.0286,
"step": 500
},
{
"epoch": 2.688172043010753,
"eval_loss": 0.032379038631916046,
"eval_runtime": 903.1767,
"eval_samples_per_second": 0.227,
"eval_steps_per_second": 0.029,
"eval_wer": 85.87243015287297,
"step": 500
},
{
"epoch": 2.8225806451612905,
"grad_norm": 115137.6015625,
"learning_rate": 6.597222222222223e-06,
"loss": 0.0293,
"step": 525
},
{
"epoch": 2.956989247311828,
"grad_norm": 148418.078125,
"learning_rate": 6.25e-06,
"loss": 0.0264,
"step": 550
},
{
"epoch": 3.0913978494623655,
"grad_norm": 44248.91796875,
"learning_rate": 5.902777777777778e-06,
"loss": 0.0157,
"step": 575
},
{
"epoch": 3.225806451612903,
"grad_norm": 86631.046875,
"learning_rate": 5.555555555555556e-06,
"loss": 0.0101,
"step": 600
},
{
"epoch": 3.360215053763441,
"grad_norm": 69759.2421875,
"learning_rate": 5.208333333333334e-06,
"loss": 0.0175,
"step": 625
},
{
"epoch": 3.4946236559139785,
"grad_norm": 79651.625,
"learning_rate": 4.861111111111111e-06,
"loss": 0.0121,
"step": 650
},
{
"epoch": 3.629032258064516,
"grad_norm": 78669.6171875,
"learning_rate": 4.513888888888889e-06,
"loss": 0.0137,
"step": 675
},
{
"epoch": 3.763440860215054,
"grad_norm": 104462.546875,
"learning_rate": 4.166666666666667e-06,
"loss": 0.0114,
"step": 700
},
{
"epoch": 3.8978494623655915,
"grad_norm": 79927.2109375,
"learning_rate": 3.819444444444445e-06,
"loss": 0.0105,
"step": 725
},
{
"epoch": 4.032258064516129,
"grad_norm": 69011.8984375,
"learning_rate": 3.4722222222222224e-06,
"loss": 0.0074,
"step": 750
},
{
"epoch": 4.166666666666667,
"grad_norm": 43131.57421875,
"learning_rate": 3.125e-06,
"loss": 0.0036,
"step": 775
},
{
"epoch": 4.301075268817204,
"grad_norm": 15227.64453125,
"learning_rate": 2.777777777777778e-06,
"loss": 0.0038,
"step": 800
},
{
"epoch": 4.435483870967742,
"grad_norm": 27650.37890625,
"learning_rate": 2.4305555555555557e-06,
"loss": 0.0082,
"step": 825
},
{
"epoch": 4.56989247311828,
"grad_norm": 146017.515625,
"learning_rate": 2.0833333333333334e-06,
"loss": 0.0066,
"step": 850
},
{
"epoch": 4.704301075268817,
"grad_norm": 41058.7421875,
"learning_rate": 1.7361111111111112e-06,
"loss": 0.0047,
"step": 875
},
{
"epoch": 4.838709677419355,
"grad_norm": 6614.59765625,
"learning_rate": 1.388888888888889e-06,
"loss": 0.0051,
"step": 900
},
{
"epoch": 4.973118279569892,
"grad_norm": 24049.263671875,
"learning_rate": 1.0416666666666667e-06,
"loss": 0.0026,
"step": 925
},
{
"epoch": 5.10752688172043,
"grad_norm": 42294.18359375,
"learning_rate": 6.944444444444445e-07,
"loss": 0.0047,
"step": 950
},
{
"epoch": 5.241935483870968,
"grad_norm": 9962.9453125,
"learning_rate": 3.4722222222222224e-07,
"loss": 0.0013,
"step": 975
},
{
"epoch": 5.376344086021505,
"grad_norm": 18759.73828125,
"learning_rate": 0.0,
"loss": 0.0047,
"step": 1000
},
{
"epoch": 5.376344086021505,
"eval_loss": 0.0025792771484702826,
"eval_runtime": 847.1283,
"eval_samples_per_second": 0.242,
"eval_steps_per_second": 0.031,
"eval_wer": 96.36267791249341,
"step": 1000
}
],
"logging_steps": 25,
"max_steps": 1000,
"num_input_tokens_seen": 0,
"num_train_epochs": 6,
"save_steps": 500,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": true
},
"attributes": {}
}
},
"total_flos": 3.4543672418304e+18,
"train_batch_size": 12,
"trial_name": null,
"trial_params": null
}