whisper-atcosim3 / trainer_state.json
luigisaetta's picture
End of training
9bb8267
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 2.3587223587223587,
"global_step": 600,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.1,
"learning_rate": 1.5333333333333334e-06,
"loss": 2.3001,
"step": 25
},
{
"epoch": 0.2,
"learning_rate": 3.2000000000000003e-06,
"loss": 0.8218,
"step": 50
},
{
"epoch": 0.2,
"eval_loss": 0.3785243332386017,
"eval_runtime": 583.9852,
"eval_samples_per_second": 3.351,
"eval_steps_per_second": 0.839,
"eval_wer": 0.14508231321232587,
"step": 50
},
{
"epoch": 0.29,
"learning_rate": 4.800000000000001e-06,
"loss": 0.2521,
"step": 75
},
{
"epoch": 0.39,
"learning_rate": 6.466666666666667e-06,
"loss": 0.1429,
"step": 100
},
{
"epoch": 0.39,
"eval_loss": 0.12126336246728897,
"eval_runtime": 580.8543,
"eval_samples_per_second": 3.369,
"eval_steps_per_second": 0.844,
"eval_wer": 0.07142254115660616,
"step": 100
},
{
"epoch": 0.49,
"learning_rate": 8.133333333333334e-06,
"loss": 0.1167,
"step": 125
},
{
"epoch": 0.59,
"learning_rate": 9.800000000000001e-06,
"loss": 0.1155,
"step": 150
},
{
"epoch": 0.59,
"eval_loss": 0.08073808252811432,
"eval_runtime": 580.5596,
"eval_samples_per_second": 3.371,
"eval_steps_per_second": 0.844,
"eval_wer": 0.05166737019839595,
"step": 150
},
{
"epoch": 0.69,
"learning_rate": 9.511111111111112e-06,
"loss": 0.0949,
"step": 175
},
{
"epoch": 0.79,
"learning_rate": 8.955555555555555e-06,
"loss": 0.0764,
"step": 200
},
{
"epoch": 0.79,
"eval_loss": 0.06518173962831497,
"eval_runtime": 562.2496,
"eval_samples_per_second": 3.481,
"eval_steps_per_second": 0.871,
"eval_wer": 0.027184466019417475,
"step": 200
},
{
"epoch": 0.88,
"learning_rate": 8.400000000000001e-06,
"loss": 0.0707,
"step": 225
},
{
"epoch": 0.98,
"learning_rate": 7.844444444444446e-06,
"loss": 0.0724,
"step": 250
},
{
"epoch": 0.98,
"eval_loss": 0.06074482947587967,
"eval_runtime": 579.9609,
"eval_samples_per_second": 3.374,
"eval_steps_per_second": 0.845,
"eval_wer": 0.03934149430139299,
"step": 250
},
{
"epoch": 1.08,
"learning_rate": 7.28888888888889e-06,
"loss": 0.0401,
"step": 275
},
{
"epoch": 1.18,
"learning_rate": 6.733333333333334e-06,
"loss": 0.0357,
"step": 300
},
{
"epoch": 1.18,
"eval_loss": 0.05690562352538109,
"eval_runtime": 561.3671,
"eval_samples_per_second": 3.486,
"eval_steps_per_second": 0.873,
"eval_wer": 0.024229632756437317,
"step": 300
},
{
"epoch": 1.28,
"learning_rate": 6.177777777777778e-06,
"loss": 0.0345,
"step": 325
},
{
"epoch": 1.38,
"learning_rate": 5.622222222222222e-06,
"loss": 0.03,
"step": 350
},
{
"epoch": 1.38,
"eval_loss": 0.055250346660614014,
"eval_runtime": 564.9558,
"eval_samples_per_second": 3.464,
"eval_steps_per_second": 0.867,
"eval_wer": 0.024314056563951036,
"step": 350
},
{
"epoch": 1.47,
"learning_rate": 5.0666666666666676e-06,
"loss": 0.0372,
"step": 375
},
{
"epoch": 1.57,
"learning_rate": 4.511111111111111e-06,
"loss": 0.0325,
"step": 400
},
{
"epoch": 1.57,
"eval_loss": 0.05560004711151123,
"eval_runtime": 557.3555,
"eval_samples_per_second": 3.511,
"eval_steps_per_second": 0.879,
"eval_wer": 0.022836639932460955,
"step": 400
},
{
"epoch": 1.67,
"learning_rate": 3.955555555555556e-06,
"loss": 0.0299,
"step": 425
},
{
"epoch": 1.77,
"learning_rate": 3.4000000000000005e-06,
"loss": 0.03,
"step": 450
},
{
"epoch": 1.77,
"eval_loss": 0.050134915858507156,
"eval_runtime": 559.7287,
"eval_samples_per_second": 3.496,
"eval_steps_per_second": 0.875,
"eval_wer": 0.024187420852680455,
"step": 450
},
{
"epoch": 1.87,
"learning_rate": 2.8444444444444446e-06,
"loss": 0.0308,
"step": 475
},
{
"epoch": 1.97,
"learning_rate": 2.2888888888888892e-06,
"loss": 0.0232,
"step": 500
},
{
"epoch": 1.97,
"eval_loss": 0.04845602065324783,
"eval_runtime": 562.6357,
"eval_samples_per_second": 3.478,
"eval_steps_per_second": 0.871,
"eval_wer": 0.020472773322076827,
"step": 500
},
{
"epoch": 2.06,
"learning_rate": 1.7333333333333336e-06,
"loss": 0.0211,
"step": 525
},
{
"epoch": 2.16,
"learning_rate": 1.1777777777777778e-06,
"loss": 0.0143,
"step": 550
},
{
"epoch": 2.16,
"eval_loss": 0.04800991341471672,
"eval_runtime": 558.3624,
"eval_samples_per_second": 3.505,
"eval_steps_per_second": 0.878,
"eval_wer": 0.01937526382439848,
"step": 550
},
{
"epoch": 2.26,
"learning_rate": 6.222222222222223e-07,
"loss": 0.013,
"step": 575
},
{
"epoch": 2.36,
"learning_rate": 6.666666666666668e-08,
"loss": 0.0105,
"step": 600
},
{
"epoch": 2.36,
"eval_loss": 0.04755961894989014,
"eval_runtime": 557.4658,
"eval_samples_per_second": 3.511,
"eval_steps_per_second": 0.879,
"eval_wer": 0.019755170958210215,
"step": 600
},
{
"epoch": 2.36,
"step": 600,
"total_flos": 1.9595614427723858e+19,
"train_loss": 0.1852580333749453,
"train_runtime": 12301.5592,
"train_samples_per_second": 1.561,
"train_steps_per_second": 0.049
}
],
"max_steps": 600,
"num_train_epochs": 3,
"total_flos": 1.9595614427723858e+19,
"trial_name": null,
"trial_params": null
}