global_batchsize_96_lr2e5 / all_results.json
sedrickkeh's picture
End of training
41f135f verified
{
"epoch": 2.994263862332696,
"total_flos": 5.620362022127534e+17,
"train_loss": 0.45750690551324824,
"train_runtime": 8551.5314,
"train_samples_per_second": 5.862,
"train_steps_per_second": 0.061
}