GaetanMichelet's picture
Model save
91b413e verified
{
"best_metric": 1.1562939882278442,
"best_model_checkpoint": "data/Mistral-7B_task-1_120-samples_config-2/checkpoint-22",
"epoch": 10.909090909090908,
"eval_steps": 500,
"global_step": 60,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.18181818181818182,
"grad_norm": 5.7528300285339355,
"learning_rate": 4.000000000000001e-06,
"loss": 2.5145,
"step": 1
},
{
"epoch": 0.36363636363636365,
"grad_norm": 5.776094913482666,
"learning_rate": 8.000000000000001e-06,
"loss": 2.2407,
"step": 2
},
{
"epoch": 0.7272727272727273,
"grad_norm": 5.344478607177734,
"learning_rate": 1.6000000000000003e-05,
"loss": 2.2934,
"step": 4
},
{
"epoch": 0.9090909090909091,
"eval_loss": 2.06683611869812,
"eval_runtime": 9.6964,
"eval_samples_per_second": 2.475,
"eval_steps_per_second": 2.475,
"step": 5
},
{
"epoch": 1.0909090909090908,
"grad_norm": 5.172861576080322,
"learning_rate": 2.4e-05,
"loss": 1.9992,
"step": 6
},
{
"epoch": 1.4545454545454546,
"grad_norm": 3.7875139713287354,
"learning_rate": 3.2000000000000005e-05,
"loss": 1.943,
"step": 8
},
{
"epoch": 1.8181818181818183,
"grad_norm": 2.55133318901062,
"learning_rate": 4e-05,
"loss": 1.592,
"step": 10
},
{
"epoch": 2.0,
"eval_loss": 1.4281545877456665,
"eval_runtime": 9.6526,
"eval_samples_per_second": 2.486,
"eval_steps_per_second": 2.486,
"step": 11
},
{
"epoch": 2.1818181818181817,
"grad_norm": 1.513327717781067,
"learning_rate": 4.8e-05,
"loss": 1.4459,
"step": 12
},
{
"epoch": 2.5454545454545454,
"grad_norm": 1.2081530094146729,
"learning_rate": 5.6000000000000006e-05,
"loss": 1.2989,
"step": 14
},
{
"epoch": 2.909090909090909,
"grad_norm": 1.313421368598938,
"learning_rate": 6.400000000000001e-05,
"loss": 1.2147,
"step": 16
},
{
"epoch": 2.909090909090909,
"eval_loss": 1.2906866073608398,
"eval_runtime": 9.6484,
"eval_samples_per_second": 2.487,
"eval_steps_per_second": 2.487,
"step": 16
},
{
"epoch": 3.2727272727272725,
"grad_norm": 1.2853072881698608,
"learning_rate": 7.2e-05,
"loss": 1.2222,
"step": 18
},
{
"epoch": 3.6363636363636362,
"grad_norm": 1.3786945343017578,
"learning_rate": 8e-05,
"loss": 1.1239,
"step": 20
},
{
"epoch": 4.0,
"grad_norm": 1.3327977657318115,
"learning_rate": 8.800000000000001e-05,
"loss": 1.0354,
"step": 22
},
{
"epoch": 4.0,
"eval_loss": 1.1562939882278442,
"eval_runtime": 9.6522,
"eval_samples_per_second": 2.486,
"eval_steps_per_second": 2.486,
"step": 22
},
{
"epoch": 4.363636363636363,
"grad_norm": 1.079969048500061,
"learning_rate": 9.6e-05,
"loss": 0.9381,
"step": 24
},
{
"epoch": 4.7272727272727275,
"grad_norm": 1.0780093669891357,
"learning_rate": 9.999512620046522e-05,
"loss": 0.907,
"step": 26
},
{
"epoch": 4.909090909090909,
"eval_loss": 1.1639357805252075,
"eval_runtime": 9.6638,
"eval_samples_per_second": 2.484,
"eval_steps_per_second": 2.484,
"step": 27
},
{
"epoch": 5.090909090909091,
"grad_norm": 1.3321374654769897,
"learning_rate": 9.995614150494293e-05,
"loss": 0.8262,
"step": 28
},
{
"epoch": 5.454545454545454,
"grad_norm": 1.1442748308181763,
"learning_rate": 9.987820251299122e-05,
"loss": 0.6567,
"step": 30
},
{
"epoch": 5.818181818181818,
"grad_norm": 1.4104325771331787,
"learning_rate": 9.976136999909156e-05,
"loss": 0.6665,
"step": 32
},
{
"epoch": 6.0,
"eval_loss": 1.2559081315994263,
"eval_runtime": 9.6492,
"eval_samples_per_second": 2.487,
"eval_steps_per_second": 2.487,
"step": 33
},
{
"epoch": 6.181818181818182,
"grad_norm": 1.3953301906585693,
"learning_rate": 9.96057350657239e-05,
"loss": 0.5627,
"step": 34
},
{
"epoch": 6.545454545454545,
"grad_norm": 1.4942609071731567,
"learning_rate": 9.941141907232765e-05,
"loss": 0.4296,
"step": 36
},
{
"epoch": 6.909090909090909,
"grad_norm": 1.9641660451889038,
"learning_rate": 9.917857354066931e-05,
"loss": 0.3785,
"step": 38
},
{
"epoch": 6.909090909090909,
"eval_loss": 1.4904040098190308,
"eval_runtime": 9.65,
"eval_samples_per_second": 2.487,
"eval_steps_per_second": 2.487,
"step": 38
},
{
"epoch": 7.2727272727272725,
"grad_norm": 1.701177954673767,
"learning_rate": 9.890738003669029e-05,
"loss": 0.2453,
"step": 40
},
{
"epoch": 7.636363636363637,
"grad_norm": 1.9415287971496582,
"learning_rate": 9.859805002892732e-05,
"loss": 0.1948,
"step": 42
},
{
"epoch": 8.0,
"grad_norm": 2.587886095046997,
"learning_rate": 9.825082472361557e-05,
"loss": 0.1927,
"step": 44
},
{
"epoch": 8.0,
"eval_loss": 1.8471574783325195,
"eval_runtime": 9.6587,
"eval_samples_per_second": 2.485,
"eval_steps_per_second": 2.485,
"step": 44
},
{
"epoch": 8.363636363636363,
"grad_norm": 1.5391498804092407,
"learning_rate": 9.786597487660337e-05,
"loss": 0.1077,
"step": 46
},
{
"epoch": 8.727272727272727,
"grad_norm": 1.8327051401138306,
"learning_rate": 9.744380058222483e-05,
"loss": 0.0701,
"step": 48
},
{
"epoch": 8.909090909090908,
"eval_loss": 2.019378900527954,
"eval_runtime": 9.657,
"eval_samples_per_second": 2.485,
"eval_steps_per_second": 2.485,
"step": 49
},
{
"epoch": 9.090909090909092,
"grad_norm": 1.920685887336731,
"learning_rate": 9.698463103929542e-05,
"loss": 0.0793,
"step": 50
},
{
"epoch": 9.454545454545455,
"grad_norm": 2.6274609565734863,
"learning_rate": 9.648882429441257e-05,
"loss": 0.0334,
"step": 52
},
{
"epoch": 9.818181818181818,
"grad_norm": 1.8720940351486206,
"learning_rate": 9.595676696276172e-05,
"loss": 0.0521,
"step": 54
},
{
"epoch": 10.0,
"eval_loss": 2.1608989238739014,
"eval_runtime": 9.6569,
"eval_samples_per_second": 2.485,
"eval_steps_per_second": 2.485,
"step": 55
},
{
"epoch": 10.181818181818182,
"grad_norm": 1.8427505493164062,
"learning_rate": 9.538887392664544e-05,
"loss": 0.0371,
"step": 56
},
{
"epoch": 10.545454545454545,
"grad_norm": 3.3887617588043213,
"learning_rate": 9.478558801197065e-05,
"loss": 0.031,
"step": 58
},
{
"epoch": 10.909090909090908,
"grad_norm": 1.1552501916885376,
"learning_rate": 9.414737964294636e-05,
"loss": 0.0379,
"step": 60
},
{
"epoch": 10.909090909090908,
"eval_loss": 2.1465117931365967,
"eval_runtime": 9.6652,
"eval_samples_per_second": 2.483,
"eval_steps_per_second": 2.483,
"step": 60
},
{
"epoch": 10.909090909090908,
"step": 60,
"total_flos": 1.4581627275444224e+16,
"train_loss": 0.7998790429672227,
"train_runtime": 1139.5176,
"train_samples_per_second": 3.861,
"train_steps_per_second": 0.219
}
],
"logging_steps": 2,
"max_steps": 250,
"num_input_tokens_seen": 0,
"num_train_epochs": 50,
"save_steps": 25,
"stateful_callbacks": {
"EarlyStoppingCallback": {
"args": {
"early_stopping_patience": 7,
"early_stopping_threshold": 0.0
},
"attributes": {
"early_stopping_patience_counter": 0
}
},
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": true
},
"attributes": {}
}
},
"total_flos": 1.4581627275444224e+16,
"train_batch_size": 1,
"trial_name": null,
"trial_params": null
}