b1_math_top_16 / all_results.json
neginr's picture
End of training
8be6fe9 verified
{
"epoch": 5.0,
"total_flos": 1.6116754450596495e+18,
"train_loss": 0.2824884724339493,
"train_runtime": 21501.2629,
"train_samples_per_second": 7.348,
"train_steps_per_second": 0.057
}