gpt2-RMT-2-mem512 / train_results.json
KotshinZ's picture
Model save
7900f86 verified
{
"total_flos": 5418484972388352.0,
"train_loss": 3.606253622488408,
"train_runtime": 424.9732,
"train_samples": 19883,
"train_samples_per_second": 48.742,
"train_steps_per_second": 1.522
}