RubricRM-4B-Judge-v2 / train_results.json
lliutianc's picture
upload curated artifacts (no checkpoint-* / logs / trainer_state) (batch 1/1)
a73248f verified
raw
history blame contribute delete
204 Bytes
{
"epoch": 2.0,
"total_flos": 247777267023872.0,
"train_loss": 0.5111368663159319,
"train_runtime": 21209.5954,
"train_samples_per_second": 3.696,
"train_steps_per_second": 0.058
}