Files
a1-stackexchange_superuser/train_results.json
ModelHub XC 7242c9cf82 初始化项目,由ModelHub XC社区提供模型
Model: DCAgent/a1-stackexchange_superuser
Source: Original Platform
2026-05-11 12:07:52 +08:00

16 lines
546 B
JSON

{
"achieved_tflops_per_gpu": 0.0027903363217319727,
"achieved_tflops_per_gpu_theoretical": 752.1148129467432,
"epoch": 7.0,
"loss_nan_ranks": 0,
"loss_rank_avg": 0.24196135997772217,
"mfu_percent": 0.00019719691319660583,
"mfu_percent_theoretical": 53.15299031425747,
"total_flos": 881525354463232.0,
"train_loss": 0.37171825520364404,
"train_runtime": 19745.0516,
"train_samples_per_second": 3.548,
"train_steps_per_second": 0.222,
"valid_targets_mean": 3242.5,
"valid_targets_min": 1195
}