332 lines
8.5 KiB
JSON
332 lines
8.5 KiB
JSON
{
|
|
"best_global_step": 600,
|
|
"best_metric": 1.4882553815841675,
|
|
"best_model_checkpoint": "./Llama-3.2-3b-ru-lora/checkpoint-600",
|
|
"epoch": 2.6938947368421053,
|
|
"eval_steps": 100,
|
|
"global_step": 800,
|
|
"is_hyper_param_search": false,
|
|
"is_local_process_zero": true,
|
|
"is_world_process_zero": true,
|
|
"log_history": [
|
|
{
|
|
"epoch": 0.08421052631578947,
|
|
"grad_norm": 0.22051513195037842,
|
|
"learning_rate": 8e-05,
|
|
"loss": 2.3452760314941408,
|
|
"step": 25
|
|
},
|
|
{
|
|
"epoch": 0.16842105263157894,
|
|
"grad_norm": 0.24215690791606903,
|
|
"learning_rate": 9.779326364692218e-05,
|
|
"loss": 1.873732452392578,
|
|
"step": 50
|
|
},
|
|
{
|
|
"epoch": 0.25263157894736843,
|
|
"grad_norm": 0.2318001091480255,
|
|
"learning_rate": 9.488966318234611e-05,
|
|
"loss": 1.7455740356445313,
|
|
"step": 75
|
|
},
|
|
{
|
|
"epoch": 0.3368421052631579,
|
|
"grad_norm": 0.22570297122001648,
|
|
"learning_rate": 9.198606271777003e-05,
|
|
"loss": 1.7174720764160156,
|
|
"step": 100
|
|
},
|
|
{
|
|
"epoch": 0.3368421052631579,
|
|
"eval_loss": 1.6638826131820679,
|
|
"eval_runtime": 41.5175,
|
|
"eval_samples_per_second": 3.011,
|
|
"eval_steps_per_second": 3.011,
|
|
"step": 100
|
|
},
|
|
{
|
|
"epoch": 0.42105263157894735,
|
|
"grad_norm": 0.2472986876964569,
|
|
"learning_rate": 8.908246225319396e-05,
|
|
"loss": 1.6817715454101563,
|
|
"step": 125
|
|
},
|
|
{
|
|
"epoch": 0.5052631578947369,
|
|
"grad_norm": 0.21928200125694275,
|
|
"learning_rate": 8.61788617886179e-05,
|
|
"loss": 1.5971102905273438,
|
|
"step": 150
|
|
},
|
|
{
|
|
"epoch": 0.5894736842105263,
|
|
"grad_norm": 0.24745388329029083,
|
|
"learning_rate": 8.327526132404182e-05,
|
|
"loss": 1.6087770080566406,
|
|
"step": 175
|
|
},
|
|
{
|
|
"epoch": 0.6736842105263158,
|
|
"grad_norm": 0.23894399404525757,
|
|
"learning_rate": 8.037166085946574e-05,
|
|
"loss": 1.623451690673828,
|
|
"step": 200
|
|
},
|
|
{
|
|
"epoch": 0.6736842105263158,
|
|
"eval_loss": 1.5778809785842896,
|
|
"eval_runtime": 41.3548,
|
|
"eval_samples_per_second": 3.023,
|
|
"eval_steps_per_second": 3.023,
|
|
"step": 200
|
|
},
|
|
{
|
|
"epoch": 0.7578947368421053,
|
|
"grad_norm": 0.24880653619766235,
|
|
"learning_rate": 7.746806039488966e-05,
|
|
"loss": 1.6210295104980468,
|
|
"step": 225
|
|
},
|
|
{
|
|
"epoch": 0.8421052631578947,
|
|
"grad_norm": 0.21366356313228607,
|
|
"learning_rate": 7.456445993031359e-05,
|
|
"loss": 1.5464874267578126,
|
|
"step": 250
|
|
},
|
|
{
|
|
"epoch": 0.9263157894736842,
|
|
"grad_norm": 0.22737279534339905,
|
|
"learning_rate": 7.166085946573751e-05,
|
|
"loss": 1.5327375793457032,
|
|
"step": 275
|
|
},
|
|
{
|
|
"epoch": 1.0101052631578948,
|
|
"grad_norm": 0.2151332050561905,
|
|
"learning_rate": 6.875725900116144e-05,
|
|
"loss": 1.5470263671875,
|
|
"step": 300
|
|
},
|
|
{
|
|
"epoch": 1.0101052631578948,
|
|
"eval_loss": 1.5347633361816406,
|
|
"eval_runtime": 41.2927,
|
|
"eval_samples_per_second": 3.027,
|
|
"eval_steps_per_second": 3.027,
|
|
"step": 300
|
|
},
|
|
{
|
|
"epoch": 1.0943157894736841,
|
|
"grad_norm": 0.24555769562721252,
|
|
"learning_rate": 6.585365853658538e-05,
|
|
"loss": 1.3567247009277343,
|
|
"step": 325
|
|
},
|
|
{
|
|
"epoch": 1.1785263157894736,
|
|
"grad_norm": 0.22149087488651276,
|
|
"learning_rate": 6.29500580720093e-05,
|
|
"loss": 1.363866424560547,
|
|
"step": 350
|
|
},
|
|
{
|
|
"epoch": 1.2627368421052632,
|
|
"grad_norm": 0.2511236071586609,
|
|
"learning_rate": 6.0046457607433216e-05,
|
|
"loss": 1.356480255126953,
|
|
"step": 375
|
|
},
|
|
{
|
|
"epoch": 1.3469473684210527,
|
|
"grad_norm": 0.27482861280441284,
|
|
"learning_rate": 5.714285714285714e-05,
|
|
"loss": 1.2976112365722656,
|
|
"step": 400
|
|
},
|
|
{
|
|
"epoch": 1.3469473684210527,
|
|
"eval_loss": 1.5319331884384155,
|
|
"eval_runtime": 41.5027,
|
|
"eval_samples_per_second": 3.012,
|
|
"eval_steps_per_second": 3.012,
|
|
"step": 400
|
|
},
|
|
{
|
|
"epoch": 1.4311578947368422,
|
|
"grad_norm": 0.3796673119068146,
|
|
"learning_rate": 5.423925667828107e-05,
|
|
"loss": 1.3674261474609375,
|
|
"step": 425
|
|
},
|
|
{
|
|
"epoch": 1.5153684210526315,
|
|
"grad_norm": 0.2730187773704529,
|
|
"learning_rate": 5.1335656213705e-05,
|
|
"loss": 1.31255615234375,
|
|
"step": 450
|
|
},
|
|
{
|
|
"epoch": 1.5995789473684212,
|
|
"grad_norm": 0.23317836225032806,
|
|
"learning_rate": 4.8432055749128926e-05,
|
|
"loss": 1.3395741271972657,
|
|
"step": 475
|
|
},
|
|
{
|
|
"epoch": 1.6837894736842105,
|
|
"grad_norm": 0.2835581302642822,
|
|
"learning_rate": 4.5528455284552844e-05,
|
|
"loss": 1.3206443786621094,
|
|
"step": 500
|
|
},
|
|
{
|
|
"epoch": 1.6837894736842105,
|
|
"eval_loss": 1.5011619329452515,
|
|
"eval_runtime": 41.4694,
|
|
"eval_samples_per_second": 3.014,
|
|
"eval_steps_per_second": 3.014,
|
|
"step": 500
|
|
},
|
|
{
|
|
"epoch": 1.768,
|
|
"grad_norm": 0.27715668082237244,
|
|
"learning_rate": 4.262485481997677e-05,
|
|
"loss": 1.3208406066894531,
|
|
"step": 525
|
|
},
|
|
{
|
|
"epoch": 1.8522105263157895,
|
|
"grad_norm": 0.27916809916496277,
|
|
"learning_rate": 3.97212543554007e-05,
|
|
"loss": 1.2577701568603517,
|
|
"step": 550
|
|
},
|
|
{
|
|
"epoch": 1.9364210526315788,
|
|
"grad_norm": 0.3093879222869873,
|
|
"learning_rate": 3.681765389082462e-05,
|
|
"loss": 1.2785945892333985,
|
|
"step": 575
|
|
},
|
|
{
|
|
"epoch": 2.0202105263157897,
|
|
"grad_norm": 0.25554776191711426,
|
|
"learning_rate": 3.391405342624855e-05,
|
|
"loss": 1.231337203979492,
|
|
"step": 600
|
|
},
|
|
{
|
|
"epoch": 2.0202105263157897,
|
|
"eval_loss": 1.4882553815841675,
|
|
"eval_runtime": 41.3478,
|
|
"eval_samples_per_second": 3.023,
|
|
"eval_steps_per_second": 3.023,
|
|
"step": 600
|
|
},
|
|
{
|
|
"epoch": 2.104421052631579,
|
|
"grad_norm": 0.2723052203655243,
|
|
"learning_rate": 3.101045296167247e-05,
|
|
"loss": 1.1024535369873047,
|
|
"step": 625
|
|
},
|
|
{
|
|
"epoch": 2.1886315789473683,
|
|
"grad_norm": 0.31523096561431885,
|
|
"learning_rate": 2.8106852497096402e-05,
|
|
"loss": 1.0629356384277344,
|
|
"step": 650
|
|
},
|
|
{
|
|
"epoch": 2.272842105263158,
|
|
"grad_norm": 0.3069048821926117,
|
|
"learning_rate": 2.5203252032520324e-05,
|
|
"loss": 1.0792010498046876,
|
|
"step": 675
|
|
},
|
|
{
|
|
"epoch": 2.3570526315789473,
|
|
"grad_norm": 0.346049964427948,
|
|
"learning_rate": 2.2299651567944253e-05,
|
|
"loss": 1.0728897094726562,
|
|
"step": 700
|
|
},
|
|
{
|
|
"epoch": 2.3570526315789473,
|
|
"eval_loss": 1.5334949493408203,
|
|
"eval_runtime": 41.3821,
|
|
"eval_samples_per_second": 3.021,
|
|
"eval_steps_per_second": 3.021,
|
|
"step": 700
|
|
},
|
|
{
|
|
"epoch": 2.441263157894737,
|
|
"grad_norm": 0.3051030933856964,
|
|
"learning_rate": 1.9396051103368175e-05,
|
|
"loss": 1.0854484558105468,
|
|
"step": 725
|
|
},
|
|
{
|
|
"epoch": 2.5254736842105263,
|
|
"grad_norm": 0.34286025166511536,
|
|
"learning_rate": 1.6492450638792104e-05,
|
|
"loss": 1.0880242919921874,
|
|
"step": 750
|
|
},
|
|
{
|
|
"epoch": 2.609684210526316,
|
|
"grad_norm": 0.30227968096733093,
|
|
"learning_rate": 1.3588850174216028e-05,
|
|
"loss": 1.0594041442871094,
|
|
"step": 775
|
|
},
|
|
{
|
|
"epoch": 2.6938947368421053,
|
|
"grad_norm": 0.3472527861595154,
|
|
"learning_rate": 1.0685249709639954e-05,
|
|
"loss": 1.0973897552490235,
|
|
"step": 800
|
|
},
|
|
{
|
|
"epoch": 2.6938947368421053,
|
|
"eval_loss": 1.5199450254440308,
|
|
"eval_runtime": 41.349,
|
|
"eval_samples_per_second": 3.023,
|
|
"eval_steps_per_second": 3.023,
|
|
"step": 800
|
|
}
|
|
],
|
|
"logging_steps": 25,
|
|
"max_steps": 891,
|
|
"num_input_tokens_seen": 0,
|
|
"num_train_epochs": 3,
|
|
"save_steps": 100,
|
|
"stateful_callbacks": {
|
|
"EarlyStoppingCallback": {
|
|
"args": {
|
|
"early_stopping_patience": 3,
|
|
"early_stopping_threshold": 0.0
|
|
},
|
|
"attributes": {
|
|
"early_stopping_patience_counter": 2
|
|
}
|
|
},
|
|
"TrainerControl": {
|
|
"args": {
|
|
"should_epoch_stop": false,
|
|
"should_evaluate": false,
|
|
"should_log": false,
|
|
"should_save": true,
|
|
"should_training_stop": false
|
|
},
|
|
"attributes": {}
|
|
}
|
|
},
|
|
"total_flos": 9.31830382399488e+16,
|
|
"train_batch_size": 1,
|
|
"trial_name": null,
|
|
"trial_params": null
|
|
}
|