Qwen2.5-1.5B-Open-R1-Distill / all_results.json
DeeLearning's picture
End of training
ea28c43 verified
raw
history blame
397 Bytes
{
"epoch": 4.0,
"eval_loss": 0.7432993650436401,
"eval_runtime": 1.9433,
"eval_samples": 100,
"eval_samples_per_second": 65.867,
"eval_steps_per_second": 2.058,
"total_flos": 308009150447616.0,
"train_loss": 0.7185798888728463,
"train_runtime": 4924.3182,
"train_samples": 16610,
"train_samples_per_second": 17.554,
"train_steps_per_second": 0.137
}