groderg's picture
Evaluation on the test set completed on 2024_10_31.
f59385d verified
raw
history blame contribute delete
232 Bytes
{
"epoch": 40.0,
"learning_rate": 1e-05,
"total_flos": 2.9601852123168e+17,
"train_loss": 0.64580397605896,
"train_runtime": 275.9938,
"train_samples_per_second": 27.175,
"train_steps_per_second": 1.087
}