Files
qwen3-8b-base-new-dpo-hh-ha…/train_results.json

9 lines
232 B
JSON
Raw Normal View History

{
"epoch": 0.999244142101285,
"total_flos": 0.0,
"train_loss": 1.1374638212250148,
"train_runtime": 2122.2138,
"train_samples": 42336,
"train_samples_per_second": 19.949,
"train_steps_per_second": 0.311
}