Qwen2.5-3B-Open-R1-Code-GRPO / train_results.json
Yukang's picture
Training in progress, step 509
1fadeb3 verified
raw
history blame
204 Bytes
{
"total_flos": 0.0,
"train_loss": 1.0097834765617772e-05,
"train_runtime": 80.5242,
"train_samples": 35735,
"train_samples_per_second": 3179.169,
"train_steps_per_second": 6.209
}