Files
Llama-3.2-1B-Instruct_SFT_s…/all_results.json

9 lines
259 B
JSON
Raw Normal View History

{
"ewc_loss": 0.0003185272216796875,
"total_flos": 2.1853937174671524e+18,
"train_loss": 1.1745690016519457,
"train_runtime": 2017.134,
"train_samples": 107517,
"train_samples_per_second": 159.906,
"train_steps_per_second": 0.312
}