d1_math_mc_llm / all_results.json
neginr's picture
End of training
c0b0c9e verified
{
"epoch": 5.0,
"total_flos": 4.947037641766863e+18,
"train_loss": 0.33841314686213425,
"train_runtime": 29065.9332,
"train_samples_per_second": 5.436,
"train_steps_per_second": 0.042
}