global_batchsize_96_lr4e5 / all_results.json
sedrickkeh's picture
End of training
d6d2f24 verified
raw
history blame
221 Bytes
{
"epoch": 2.994263862332696,
"total_flos": 5.620362022127534e+17,
"train_loss": 0.4276364711452261,
"train_runtime": 8560.5176,
"train_samples_per_second": 5.856,
"train_steps_per_second": 0.061
}