Qwen2.5-0.5B-Open-R1-Distill / all_results.json
herman66's picture
End of training
29c8195 verified
raw
history blame contribute delete
381 Bytes
{
"eval_loss": 1.1877331733703613,
"eval_runtime": 39.9222,
"eval_samples": 100,
"eval_samples_per_second": 25.65,
"eval_steps_per_second": 6.412,
"total_flos": 1.900575720430633e+17,
"train_loss": 1.19111062203986,
"train_runtime": 26492.3013,
"train_samples": 16610,
"train_samples_per_second": 6.525,
"train_steps_per_second": 0.408
}