Qwen2.5-0.5B-Open-R1-SFT / train_results.json
DGME's picture
Model save
cd8544d verified
{
"total_flos": 4.719554866967808e+16,
"train_loss": 0.0366410073112039,
"train_runtime": 308.342,
"train_samples": 7473,
"train_samples_per_second": 4.573,
"train_steps_per_second": 0.551
}