train_math_qa_42_1760637608 / train_results.json
rbelanec's picture
End of training
ab6d87a verified
{
"epoch": 20.0,
"num_input_tokens_seen": 77902976,
"total_flos": 3.5080602449577247e+18,
"train_loss": 1.0952728209393985,
"train_runtime": 26067.2795,
"train_samples_per_second": 20.603,
"train_steps_per_second": 5.151
}