b1_math_top_4_1k / train_results.json
ryanmarten's picture
Upload model
7a1f0f5 verified
{
"epoch": 6.72,
"total_flos": 1.1913104726582886e+17,
"train_loss": 0.4864661923476628,
"train_runtime": 5735.7432,
"train_samples_per_second": 1.22,
"train_steps_per_second": 0.012
}