Files
cds_sh1_fr_30/checkpoint-290/trainer_state.json

72 lines
1.7 KiB
JSON
Raw Normal View History

{
"best_global_step": 290,
"best_metric": 4.028092384338379,
"best_model_checkpoint": "/home/p318482/syntactic-bootstrapping/model_trained/cds_sh1_fr_30/checkpoint-290",
"epoch": 2.0,
"eval_steps": 500,
"global_step": 290,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.006896551724137931,
"grad_norm": 34.13589859008789,
"learning_rate": 0.0,
"loss": 9.8164,
"step": 1
},
{
"epoch": 1.0,
"grad_norm": 1.4214282035827637,
"learning_rate": 2.88e-05,
"loss": 6.0017,
"step": 145
},
{
"epoch": 1.0,
"eval_loss": 4.516501426696777,
"eval_runtime": 4.3416,
"eval_samples_per_second": 810.983,
"eval_steps_per_second": 3.225,
"step": 145
},
{
"epoch": 2.0,
"grad_norm": 1.7844622135162354,
"learning_rate": 5.7799999999999995e-05,
"loss": 4.1747,
"step": 290
},
{
"epoch": 2.0,
"eval_loss": 4.028092384338379,
"eval_runtime": 4.4671,
"eval_samples_per_second": 788.2,
"eval_steps_per_second": 3.134,
"step": 290
}
],
"logging_steps": 500,
"max_steps": 2900,
"num_input_tokens_seen": 0,
"num_train_epochs": 20,
"save_steps": 500,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": false
},
"attributes": {}
}
},
"total_flos": 4824888016896000.0,
"train_batch_size": 256,
"trial_name": null,
"trial_params": null
}