b1_math_top_16_3k / train_results.json
ryanmarten's picture
Upload model
0d6ec42 verified
{
"epoch": 6.850632911392405,
"total_flos": 2.86943794002133e+17,
"train_loss": 0.2993284181159522,
"train_runtime": 12424.2407,
"train_samples_per_second": 1.78,
"train_steps_per_second": 0.018
}