global_batchsize_512_lradjusted8 / train_results.json
sedrickkeh's picture
End of training
85fafc8 verified
raw
history blame contribute delete
223 Bytes
{
"epoch": 2.9943851768669285,
"total_flos": 7.06867120956244e+18,
"train_loss": 0.4220664166025929,
"train_runtime": 109600.6183,
"train_samples_per_second": 3.119,
"train_steps_per_second": 0.006
}