tielectra-geezswitch / trainer_state.json
fgaim's picture
Config and tokenizer
86f04b9
raw
history blame
710 Bytes
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 3.0,
"global_step": 705,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 2.13,
"learning_rate": 5.815602836879432e-06,
"loss": 0.7527,
"step": 500
},
{
"epoch": 3.0,
"step": 705,
"total_flos": 165498802560000.0,
"train_loss": 0.6064913972895196,
"train_runtime": 192.036,
"train_samples_per_second": 117.166,
"train_steps_per_second": 3.671
}
],
"max_steps": 705,
"num_train_epochs": 3,
"total_flos": 165498802560000.0,
"trial_name": null,
"trial_params": null
}