Qwen2.5-Math-1.5B-Open-R1-Distill / train_results.json
a-F1's picture
Model save
1c5b27e verified
{
"total_flos": 6.960985517882081e+17,
"train_loss": 0.8181797522337325,
"train_runtime": 16039.8645,
"train_samples": 16610,
"train_samples_per_second": 1.348,
"train_steps_per_second": 0.168
}