Model: Neelectric/Llama-3.1-8B-Instruct_SFT_sciencefisher_v00.06 Source: Original Platform
9 lines
260 B
JSON
9 lines
260 B
JSON
{
|
|
"ewc_loss": 0.0002918243408203125,
|
|
"total_flos": 3.1829322264618205e+19,
|
|
"train_loss": 0.9272488037147323,
|
|
"train_runtime": 37031.0594,
|
|
"train_samples": 145693,
|
|
"train_samples_per_second": 11.803,
|
|
"train_steps_per_second": 0.738
|
|
} |