359 lines
8.4 KiB
JSON
359 lines
8.4 KiB
JSON
{
|
|
"best_metric": 17.5811209439528,
|
|
"best_model_checkpoint": "./whisper-small-ablation-add-NLD-90/checkpoint-1000",
|
|
"epoch": 2.6315789473684212,
|
|
"eval_steps": 200,
|
|
"global_step": 1000,
|
|
"is_hyper_param_search": false,
|
|
"is_local_process_zero": true,
|
|
"is_world_process_zero": true,
|
|
"log_history": [
|
|
{
|
|
"epoch": 0.06578947368421052,
|
|
"grad_norm": 21.71042823791504,
|
|
"learning_rate": 5.000000000000001e-07,
|
|
"loss": 1.538,
|
|
"step": 25
|
|
},
|
|
{
|
|
"epoch": 0.13157894736842105,
|
|
"grad_norm": 17.377655029296875,
|
|
"learning_rate": 1.0000000000000002e-06,
|
|
"loss": 1.2262,
|
|
"step": 50
|
|
},
|
|
{
|
|
"epoch": 0.19736842105263158,
|
|
"grad_norm": 7.760909080505371,
|
|
"learning_rate": 1.5e-06,
|
|
"loss": 0.7052,
|
|
"step": 75
|
|
},
|
|
{
|
|
"epoch": 0.2631578947368421,
|
|
"grad_norm": 6.189646244049072,
|
|
"learning_rate": 2.0000000000000003e-06,
|
|
"loss": 0.4282,
|
|
"step": 100
|
|
},
|
|
{
|
|
"epoch": 0.32894736842105265,
|
|
"grad_norm": 4.806670188903809,
|
|
"learning_rate": 2.5e-06,
|
|
"loss": 0.3623,
|
|
"step": 125
|
|
},
|
|
{
|
|
"epoch": 0.39473684210526316,
|
|
"grad_norm": 5.8449554443359375,
|
|
"learning_rate": 3e-06,
|
|
"loss": 0.349,
|
|
"step": 150
|
|
},
|
|
{
|
|
"epoch": 0.4605263157894737,
|
|
"grad_norm": 6.015202522277832,
|
|
"learning_rate": 3.5e-06,
|
|
"loss": 0.3177,
|
|
"step": 175
|
|
},
|
|
{
|
|
"epoch": 0.5263157894736842,
|
|
"grad_norm": 5.104579925537109,
|
|
"learning_rate": 4.000000000000001e-06,
|
|
"loss": 0.2897,
|
|
"step": 200
|
|
},
|
|
{
|
|
"epoch": 0.5263157894736842,
|
|
"eval_loss": 0.42691341042518616,
|
|
"eval_runtime": 104.1539,
|
|
"eval_samples_per_second": 6.596,
|
|
"eval_steps_per_second": 0.413,
|
|
"eval_wer": 18.672566371681416,
|
|
"step": 200
|
|
},
|
|
{
|
|
"epoch": 0.5921052631578947,
|
|
"grad_norm": 4.69817590713501,
|
|
"learning_rate": 4.5e-06,
|
|
"loss": 0.3183,
|
|
"step": 225
|
|
},
|
|
{
|
|
"epoch": 0.6578947368421053,
|
|
"grad_norm": 6.035786151885986,
|
|
"learning_rate": 5e-06,
|
|
"loss": 0.2797,
|
|
"step": 250
|
|
},
|
|
{
|
|
"epoch": 0.7236842105263158,
|
|
"grad_norm": 5.456104755401611,
|
|
"learning_rate": 5.500000000000001e-06,
|
|
"loss": 0.2788,
|
|
"step": 275
|
|
},
|
|
{
|
|
"epoch": 0.7894736842105263,
|
|
"grad_norm": 6.070672988891602,
|
|
"learning_rate": 6e-06,
|
|
"loss": 0.2551,
|
|
"step": 300
|
|
},
|
|
{
|
|
"epoch": 0.8552631578947368,
|
|
"grad_norm": 5.564907073974609,
|
|
"learning_rate": 6.5000000000000004e-06,
|
|
"loss": 0.2558,
|
|
"step": 325
|
|
},
|
|
{
|
|
"epoch": 0.9210526315789473,
|
|
"grad_norm": 5.393413066864014,
|
|
"learning_rate": 7e-06,
|
|
"loss": 0.2507,
|
|
"step": 350
|
|
},
|
|
{
|
|
"epoch": 0.9868421052631579,
|
|
"grad_norm": 4.764667987823486,
|
|
"learning_rate": 7.500000000000001e-06,
|
|
"loss": 0.2429,
|
|
"step": 375
|
|
},
|
|
{
|
|
"epoch": 1.0526315789473684,
|
|
"grad_norm": 4.35653018951416,
|
|
"learning_rate": 8.000000000000001e-06,
|
|
"loss": 0.177,
|
|
"step": 400
|
|
},
|
|
{
|
|
"epoch": 1.0526315789473684,
|
|
"eval_loss": 0.41943758726119995,
|
|
"eval_runtime": 105.1972,
|
|
"eval_samples_per_second": 6.531,
|
|
"eval_steps_per_second": 0.409,
|
|
"eval_wer": 18.08259587020649,
|
|
"step": 400
|
|
},
|
|
{
|
|
"epoch": 1.118421052631579,
|
|
"grad_norm": 4.236116409301758,
|
|
"learning_rate": 8.5e-06,
|
|
"loss": 0.1628,
|
|
"step": 425
|
|
},
|
|
{
|
|
"epoch": 1.1842105263157894,
|
|
"grad_norm": 4.181164264678955,
|
|
"learning_rate": 9e-06,
|
|
"loss": 0.1728,
|
|
"step": 450
|
|
},
|
|
{
|
|
"epoch": 1.25,
|
|
"grad_norm": 4.1720452308654785,
|
|
"learning_rate": 9.5e-06,
|
|
"loss": 0.1684,
|
|
"step": 475
|
|
},
|
|
{
|
|
"epoch": 1.3157894736842106,
|
|
"grad_norm": 4.402536392211914,
|
|
"learning_rate": 1e-05,
|
|
"loss": 0.1662,
|
|
"step": 500
|
|
},
|
|
{
|
|
"epoch": 1.381578947368421,
|
|
"grad_norm": 4.496105194091797,
|
|
"learning_rate": 9.5e-06,
|
|
"loss": 0.1589,
|
|
"step": 525
|
|
},
|
|
{
|
|
"epoch": 1.4473684210526316,
|
|
"grad_norm": 5.0732598304748535,
|
|
"learning_rate": 9e-06,
|
|
"loss": 0.1757,
|
|
"step": 550
|
|
},
|
|
{
|
|
"epoch": 1.513157894736842,
|
|
"grad_norm": 4.759901523590088,
|
|
"learning_rate": 8.5e-06,
|
|
"loss": 0.16,
|
|
"step": 575
|
|
},
|
|
{
|
|
"epoch": 1.5789473684210527,
|
|
"grad_norm": 4.546085357666016,
|
|
"learning_rate": 8.000000000000001e-06,
|
|
"loss": 0.1516,
|
|
"step": 600
|
|
},
|
|
{
|
|
"epoch": 1.5789473684210527,
|
|
"eval_loss": 0.4376579523086548,
|
|
"eval_runtime": 104.9438,
|
|
"eval_samples_per_second": 6.546,
|
|
"eval_steps_per_second": 0.41,
|
|
"eval_wer": 17.97197640117994,
|
|
"step": 600
|
|
},
|
|
{
|
|
"epoch": 1.6447368421052633,
|
|
"grad_norm": 4.176891326904297,
|
|
"learning_rate": 7.500000000000001e-06,
|
|
"loss": 0.1631,
|
|
"step": 625
|
|
},
|
|
{
|
|
"epoch": 1.7105263157894737,
|
|
"grad_norm": 4.341442584991455,
|
|
"learning_rate": 7e-06,
|
|
"loss": 0.1645,
|
|
"step": 650
|
|
},
|
|
{
|
|
"epoch": 1.776315789473684,
|
|
"grad_norm": 3.9501426219940186,
|
|
"learning_rate": 6.5000000000000004e-06,
|
|
"loss": 0.155,
|
|
"step": 675
|
|
},
|
|
{
|
|
"epoch": 1.8421052631578947,
|
|
"grad_norm": 3.759936809539795,
|
|
"learning_rate": 6e-06,
|
|
"loss": 0.156,
|
|
"step": 700
|
|
},
|
|
{
|
|
"epoch": 1.9078947368421053,
|
|
"grad_norm": 3.296854257583618,
|
|
"learning_rate": 5.500000000000001e-06,
|
|
"loss": 0.1545,
|
|
"step": 725
|
|
},
|
|
{
|
|
"epoch": 1.973684210526316,
|
|
"grad_norm": 4.553784370422363,
|
|
"learning_rate": 5e-06,
|
|
"loss": 0.1481,
|
|
"step": 750
|
|
},
|
|
{
|
|
"epoch": 2.039473684210526,
|
|
"grad_norm": 1.8281787633895874,
|
|
"learning_rate": 4.5e-06,
|
|
"loss": 0.1124,
|
|
"step": 775
|
|
},
|
|
{
|
|
"epoch": 2.1052631578947367,
|
|
"grad_norm": 2.578198194503784,
|
|
"learning_rate": 4.000000000000001e-06,
|
|
"loss": 0.0713,
|
|
"step": 800
|
|
},
|
|
{
|
|
"epoch": 2.1052631578947367,
|
|
"eval_loss": 0.4024522602558136,
|
|
"eval_runtime": 104.903,
|
|
"eval_samples_per_second": 6.549,
|
|
"eval_steps_per_second": 0.41,
|
|
"eval_wer": 17.831858407079647,
|
|
"step": 800
|
|
},
|
|
{
|
|
"epoch": 2.1710526315789473,
|
|
"grad_norm": 2.781614065170288,
|
|
"learning_rate": 3.5e-06,
|
|
"loss": 0.0723,
|
|
"step": 825
|
|
},
|
|
{
|
|
"epoch": 2.236842105263158,
|
|
"grad_norm": 3.354801654815674,
|
|
"learning_rate": 3e-06,
|
|
"loss": 0.0653,
|
|
"step": 850
|
|
},
|
|
{
|
|
"epoch": 2.3026315789473686,
|
|
"grad_norm": 2.1840062141418457,
|
|
"learning_rate": 2.5e-06,
|
|
"loss": 0.0729,
|
|
"step": 875
|
|
},
|
|
{
|
|
"epoch": 2.3684210526315788,
|
|
"grad_norm": 1.947718858718872,
|
|
"learning_rate": 2.0000000000000003e-06,
|
|
"loss": 0.0752,
|
|
"step": 900
|
|
},
|
|
{
|
|
"epoch": 2.4342105263157894,
|
|
"grad_norm": 1.8653781414031982,
|
|
"learning_rate": 1.5e-06,
|
|
"loss": 0.0831,
|
|
"step": 925
|
|
},
|
|
{
|
|
"epoch": 2.5,
|
|
"grad_norm": 1.6506271362304688,
|
|
"learning_rate": 1.0000000000000002e-06,
|
|
"loss": 0.0695,
|
|
"step": 950
|
|
},
|
|
{
|
|
"epoch": 2.5657894736842106,
|
|
"grad_norm": 2.864332437515259,
|
|
"learning_rate": 5.000000000000001e-07,
|
|
"loss": 0.0686,
|
|
"step": 975
|
|
},
|
|
{
|
|
"epoch": 2.6315789473684212,
|
|
"grad_norm": 2.2327263355255127,
|
|
"learning_rate": 0.0,
|
|
"loss": 0.074,
|
|
"step": 1000
|
|
},
|
|
{
|
|
"epoch": 2.6315789473684212,
|
|
"eval_loss": 0.425365149974823,
|
|
"eval_runtime": 104.7256,
|
|
"eval_samples_per_second": 6.56,
|
|
"eval_steps_per_second": 0.411,
|
|
"eval_wer": 17.5811209439528,
|
|
"step": 1000
|
|
}
|
|
],
|
|
"logging_steps": 25,
|
|
"max_steps": 1000,
|
|
"num_input_tokens_seen": 0,
|
|
"num_train_epochs": 3,
|
|
"save_steps": 200,
|
|
"stateful_callbacks": {
|
|
"TrainerControl": {
|
|
"args": {
|
|
"should_epoch_stop": false,
|
|
"should_evaluate": false,
|
|
"should_log": false,
|
|
"should_save": true,
|
|
"should_training_stop": true
|
|
},
|
|
"attributes": {}
|
|
}
|
|
},
|
|
"total_flos": 9.220303540224e+18,
|
|
"train_batch_size": 32,
|
|
"trial_name": null,
|
|
"trial_params": null
|
|
}
|