1SV72 / train_results.json
gotzmann's picture
..
f1faee6
{
"epoch": 1.0,
"total_flos": 1.7000184777585721e+19,
"train_loss": 1.553016853588025,
"train_runtime": 28326.3102,
"train_samples_per_second": 0.843,
"train_steps_per_second": 0.105
}