mcmaster-llama3-1b-full-pt / train_results.json
keatone's picture
End of training
ede525c verified
raw
history blame contribute delete
221 Bytes
{
"epoch": 4.998698642870422,
"total_flos": 5.14290499398402e+18,
"train_loss": 0.19145491501161208,
"train_runtime": 31931.328,
"train_samples_per_second": 6.737,
"train_steps_per_second": 0.105
}