d1_math_long_paragraphs / train_results.json
neginr's picture
End of training
08a31e0 verified
raw
history blame contribute delete
221 Bytes
{
"epoch": 4.99486125385406,
"total_flos": 4.757804886857613e+18,
"train_loss": 0.34405366748939326,
"train_runtime": 28473.715,
"train_samples_per_second": 5.463,
"train_steps_per_second": 0.043
}