b2_math_random_0.3k / train_results.json
ryanmarten's picture
End of training
6b0de72 verified
{
"epoch": 12.30379746835443,
"total_flos": 5.484038039850189e+16,
"train_loss": 0.2492905456540931,
"train_runtime": 3417.7179,
"train_samples_per_second": 1.202,
"train_steps_per_second": 0.034
}