a1_math_gair_math / train_results.json
gsmyrnis's picture
End of training
b9adc88 verified
{
"epoch": 5.0,
"total_flos": 2564837740773376.0,
"train_loss": 0.45295593604626444,
"train_runtime": 27969.5706,
"train_samples_per_second": 5.649,
"train_steps_per_second": 0.044
}