robertuitonews-tweetcontext / eval_results.json
finiteautomata's picture
First model
259d68a
{
"epoch": 4.71,
"eval_loss": 1.5495115518569946,
"eval_runtime": 200.1691,
"eval_samples": 181300,
"eval_samples_per_second": 905.734,
"eval_steps_per_second": 3.542,
"perplexity": 4.709169435672752
}