qwen2-0.5b-sft / all_results.json
yangzhao02's picture
End of training
1ac92ea verified
raw
history blame
416 Bytes
{
"epoch": 0.999738425320429,
"eval_loss": 1.5079056024551392,
"eval_runtime": 417.8091,
"eval_samples": 23109,
"eval_samples_per_second": 64.802,
"eval_steps_per_second": 4.052,
"total_flos": 107512363745280.0,
"train_loss": 1.5265680355653908,
"train_runtime": 15553.3872,
"train_samples": 207864,
"train_samples_per_second": 15.73,
"train_steps_per_second": 0.123
}