e1_math_ms_phi / train_results.json
neginr's picture
End of training
cf92199 verified
raw
history blame contribute delete
224 Bytes
{
"epoch": 4.982278481012658,
"total_flos": 1.9490496700092252e+18,
"train_loss": 0.9533841308054886,
"train_runtime": 13422.2897,
"train_samples_per_second": 11.771,
"train_steps_per_second": 0.092
}