Files
llama3.2-1b-deita-dpo-stude…/train_results.json

9 lines
230 B
JSON
Raw Permalink Normal View History

{
"epoch": 3.0,
"total_flos": 90953314467840.0,
"train_loss": 0.9790556504583052,
"train_runtime": 12705.6497,
"train_samples": 9500,
"train_samples_per_second": 8.618,
"train_steps_per_second": 0.135
}