babylm-default_seed-42_1e-3 / train_results.json
qing-yao's picture
Model save
f6afe72 verified
raw
history blame contribute delete
252 Bytes
{
"epoch": 19.99718982717437,
"total_flos": 1.18991215558656e+18,
"train_loss": 3.2179960425614373,
"train_runtime": 62492.0089,
"train_samples": 455458,
"train_samples_per_second": 145.765,
"train_steps_per_second": 0.569
}