PELM-JointGPT / train_results.json
GItaf's picture
End of training
1961feb
raw
history blame
205 Bytes
{
"epoch": 3.0,
"total_flos": 1.088098614116352e+16,
"train_loss": 4.11131799351585,
"train_runtime": 5817.7923,
"train_samples_per_second": 3.579,
"train_steps_per_second": 1.789
}