d1_science_mc_llm_3k / train_results.json
ryanmarten's picture
Upload model
e100fd5 verified
{
"epoch": 7.0,
"total_flos": 3.354167586660024e+17,
"train_loss": 0.3882459873309383,
"train_runtime": 6538.9235,
"train_samples_per_second": 3.383,
"train_steps_per_second": 0.035
}