Files
Qwen3-8B_exp_tas_tmux_large…/all_results.json

16 lines
547 B
JSON
Raw Normal View History

{
"achieved_tflops_per_gpu": 0.0011402087434851426,
"achieved_tflops_per_gpu_theoretical": 226.98960570915227,
"epoch": 8.0,
"loss_nan_ranks": 0,
"loss_rank_avg": 0.016571061685681343,
"mfu_percent": 0.0003654515203478021,
"mfu_percent_theoretical": 72.75307875293342,
"total_flos": 1378268975529984.0,
"train_loss": 0.16813087239582386,
"train_runtime": 37774.5792,
"train_samples_per_second": 2.146,
"train_steps_per_second": 0.067,
"valid_targets_mean": 4493.3,
"valid_targets_min": 905
}