lama3_lora3 / checkpoint-100 /trainer_state.json
王涛
hr-train
965e32a
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 6.451612903225806,
"eval_steps": 500,
"global_step": 100,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.6451612903225806,
"grad_norm": 1.17766535282135,
"learning_rate": 9.666666666666667e-05,
"loss": 2.5686,
"step": 10
},
{
"epoch": 1.2903225806451613,
"grad_norm": 1.074232816696167,
"learning_rate": 9.333333333333334e-05,
"loss": 2.2897,
"step": 20
},
{
"epoch": 1.935483870967742,
"grad_norm": 1.3295201063156128,
"learning_rate": 9e-05,
"loss": 2.0286,
"step": 30
},
{
"epoch": 2.5806451612903225,
"grad_norm": 1.5712580680847168,
"learning_rate": 8.666666666666667e-05,
"loss": 1.6861,
"step": 40
},
{
"epoch": 3.225806451612903,
"grad_norm": 2.046724319458008,
"learning_rate": 8.333333333333334e-05,
"loss": 1.5038,
"step": 50
},
{
"epoch": 3.870967741935484,
"grad_norm": 2.154179334640503,
"learning_rate": 8e-05,
"loss": 1.1607,
"step": 60
},
{
"epoch": 4.516129032258064,
"grad_norm": 1.9853581190109253,
"learning_rate": 7.666666666666667e-05,
"loss": 0.8446,
"step": 70
},
{
"epoch": 5.161290322580645,
"grad_norm": 1.9510444402694702,
"learning_rate": 7.333333333333333e-05,
"loss": 0.6355,
"step": 80
},
{
"epoch": 5.806451612903226,
"grad_norm": 2.4067842960357666,
"learning_rate": 7e-05,
"loss": 0.396,
"step": 90
},
{
"epoch": 6.451612903225806,
"grad_norm": 2.007054090499878,
"learning_rate": 6.666666666666667e-05,
"loss": 0.2548,
"step": 100
}
],
"logging_steps": 10,
"max_steps": 300,
"num_input_tokens_seen": 0,
"num_train_epochs": 20,
"save_steps": 100,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": false
},
"attributes": {}
}
},
"total_flos": 2.89031048921088e+16,
"train_batch_size": 4,
"trial_name": null,
"trial_params": null
}