Qwen2.5-3B-Open-R1-GRPO / train_results.json
howardzhou's picture
Model save
276ab41 verified
raw
history blame contribute delete
202 Bytes
{
"total_flos": 0.0,
"train_loss": 0.004591864831248247,
"train_runtime": 81346.3601,
"train_samples": 72441,
"train_samples_per_second": 0.891,
"train_steps_per_second": 0.008
}