Model: Neelectric/Llama-3.1-8B-Instruct_SFT_sciencefisher_v00.08 Source: Original Platform
9 lines
258 B
JSON
9 lines
258 B
JSON
{
|
|
"ewc_loss": 0.00017261505126953125,
|
|
"total_flos": 3.1829322264618205e+19,
|
|
"train_loss": 0.9366044780482606,
|
|
"train_runtime": 37081.4,
|
|
"train_samples": 145693,
|
|
"train_samples_per_second": 11.787,
|
|
"train_steps_per_second": 0.737
|
|
} |