d1_science_shortest_10k / all_results.json
ryanmarten's picture
Upload model
3febca3 verified
{
"epoch": 4.992,
"total_flos": 6.670095881926083e+17,
"train_loss": 0.39215392955602746,
"train_runtime": 12456.6735,
"train_samples_per_second": 4.014,
"train_steps_per_second": 0.031
}