train_multirc_1745950259 / train_results.json
rbelanec's picture
End of training
c2420db verified
{
"epoch": 6.525328330206379,
"num_input_tokens_seen": 76963024,
"total_flos": 3.222931275084472e+17,
"train_loss": 0.1829463173724711,
"train_runtime": 28519.5689,
"train_samples_per_second": 5.61,
"train_steps_per_second": 1.403
}