Qwen2.5-3B-Open-R1-GRPO / train_results.json
dadadar's picture
Model save
1a8d880 verified
raw
history blame
200 Bytes
{
"total_flos": 0.0,
"train_loss": 162.96194878991443,
"train_runtime": 154205.9094,
"train_samples": 20000,
"train_samples_per_second": 0.13,
"train_steps_per_second": 0.016
}