Files
OLMo-2-1124-7B-Instruct_SFT…/train_results.json

11 lines
335 B
JSON
Raw Permalink Normal View History

{
"ewc_loss": 0.119140625,
"ewc_loss_diag": 0.00030517578125,
"ewc_loss_parallel": 8.869171142578125e-05,
"total_flos": 5.156325133756924e+19,
"train_loss": 0.5050858836974098,
"train_runtime": 50355.4873,
"train_samples": 125770,
"train_samples_per_second": 7.493,
"train_steps_per_second": 0.468
}