e1_math_ms_qwq / train_results.json
neginr's picture
End of training
95e5ef0 verified
raw
history blame contribute delete
223 Bytes
{
"epoch": 4.982278481012658,
"total_flos": 1.899139087207424e+18,
"train_loss": 0.5717073486345571,
"train_runtime": 13070.0674,
"train_samples_per_second": 12.089,
"train_steps_per_second": 0.094
}