llama3-sudo-sanity / train_results.json
Qin Liu
Model save
c5196f3 verified
raw
history blame
246 Bytes
{
"epoch": 2.9696969696969697,
"total_flos": 706816481427456.0,
"train_loss": 1.9096906704156578,
"train_runtime": 1619.2688,
"train_samples": 6321,
"train_samples_per_second": 11.711,
"train_steps_per_second": 0.091
}