Files
openthoughts-4-code-qwen3-3…/all_results.json

16 lines
549 B
JSON
Raw Permalink Normal View History

{
"achieved_tflops_per_gpu": 0.006294075316425587,
"achieved_tflops_per_gpu_theoretical": 530.5945130815979,
"epoch": 2.0,
"loss_nan_ranks": 0,
"loss_rank_avg": 0.598960816860199,
"mfu_percent": 0.002017331832187688,
"mfu_percent_theoretical": 170.06234393640958,
"total_flos": 2.3824840797454336e+16,
"train_loss": 0.11019459065355108,
"train_runtime": 14786.2521,
"train_samples_per_second": 129.744,
"train_steps_per_second": 0.507,
"valid_targets_mean": 15636.4,
"valid_targets_min": 509
}