phi-2-gpo-test-longest-iter-0 / train_results.json
BraylonDash's picture
Model save
26d21df verified
raw
history blame
195 Bytes
{
"epoch": 1.33,
"train_loss": 0.0029907251009717584,
"train_runtime": 7.9772,
"train_samples": 30567,
"train_samples_per_second": 2.256,
"train_steps_per_second": 0.251
}