split_context / all_results.json
tangken333's picture
End of training
203ee39 verified
raw
history blame
210 Bytes
{
"epoch": 3.0,
"total_flos": 2.2819812112387277e+17,
"train_loss": 0.38972266467027766,
"train_runtime": 34674.5164,
"train_samples_per_second": 0.386,
"train_steps_per_second": 0.008
}