100M_low_100_1208 / eval_results.json
craa's picture
End of training
13db2b4 verified
raw
history blame contribute delete
268 Bytes
{
"epoch": 10.0,
"eval_accuracy": 0.3938669356332687,
"eval_loss": 3.3093740940093994,
"eval_runtime": 210.502,
"eval_samples": 18011,
"eval_samples_per_second": 85.562,
"eval_steps_per_second": 5.349,
"perplexity": 27.367990320873247
}