train_math_qa_123_1760637721 / eval_results.json
rbelanec's picture
End of training
107f0f2 verified
{
"epoch": 20.0,
"eval_loss": 0.7989339828491211,
"eval_runtime": 70.5638,
"eval_samples_per_second": 42.288,
"eval_steps_per_second": 10.572,
"num_input_tokens_seen": 77961608
}