bart-large-scientific-papers / train_results.json
parteeksj's picture
bart-large fine tuned on the scientific papers dataset
99c6c0f
raw
history blame
195 Bytes
{
"epoch": 10.0,
"train_loss": 1.3997784228157961,
"train_runtime": 9169.8515,
"train_samples": 19706,
"train_samples_per_second": 21.49,
"train_steps_per_second": 3.582
}