train_record_42_1761957644 / eval_results.json
rbelanec's picture
End of training
b85776e verified
{
"epoch": 20.0,
"eval_loss": 0.3110341727733612,
"eval_runtime": 319.2299,
"eval_samples_per_second": 43.498,
"eval_steps_per_second": 10.876,
"num_input_tokens_seen": 929295680
}