gpt_train_12_384_new / train_results.json
gokulsrinivasagan's picture
End of training
b1b0269 verified
raw
history blame
254 Bytes
{
"epoch": 6.7809745229100065,
"total_flos": 5.861064073050849e+17,
"train_loss": 4.399306026785714,
"train_runtime": 93598.1238,
"train_samples": 660643,
"train_samples_per_second": 705.829,
"train_steps_per_second": 22.058
}