aam-len3-bs256-lr1e-3 / train_results.json
yangwang825's picture
End of training
4da0132 verified
raw
history blame contribute delete
206 Bytes
{
"epoch": 10.0,
"total_flos": 8.1484088684544e+18,
"train_loss": 1.9844060945237343,
"train_runtime": 19376.6921,
"train_samples_per_second": 69.04,
"train_steps_per_second": 0.27
}