Files
oh-dcft-v3.1-llama-3.3-70b-…/train_results.json

8 lines
219 B
JSON
Raw Normal View History

{
"epoch": 2.999194198227236,
"total_flos": 2926650442383360.0,
"train_loss": 0.3617401811868598,
"train_runtime": 75547.7589,
"train_samples_per_second": 9.461,
"train_steps_per_second": 0.074
}