qwen2-0.5b-sft / all_results.json
yangzhao02's picture
End of training
a02a0c0 verified
raw
history blame contribute delete
415 Bytes
{
"epoch": 0.9998676022772408,
"eval_loss": 1.526898980140686,
"eval_runtime": 437.3232,
"eval_samples": 23109,
"eval_samples_per_second": 61.154,
"eval_steps_per_second": 3.823,
"total_flos": 106218135748608.0,
"train_loss": 1.5447227579809852,
"train_runtime": 15813.94,
"train_samples": 207864,
"train_samples_per_second": 15.283,
"train_steps_per_second": 0.119
}