smolm-autoreg-bpe-seed_888 / eval_results.json
kanishka's picture
End of training
92997d0
raw
history blame
267 Bytes
{
"epoch": 10.0,
"eval_accuracy": 0.4861165901356519,
"eval_loss": 2.557232141494751,
"eval_runtime": 4.0987,
"eval_samples": 4491,
"eval_samples_per_second": 1095.703,
"eval_steps_per_second": 2.196,
"perplexity": 12.900062308790385
}