train_math_qa_42_1760637605 / eval_results.json
rbelanec's picture
End of training
187656a verified
{
"epoch": 20.0,
"eval_loss": 0.8015024065971375,
"eval_runtime": 71.3055,
"eval_samples_per_second": 41.848,
"eval_steps_per_second": 10.462,
"num_input_tokens_seen": 77902976
}