Files
a1-stackexchange_codereview/train_results.json
ModelHub XC 36be9d6afc 初始化项目,由ModelHub XC社区提供模型
Model: DCAgent/a1-stackexchange_codereview
Source: Original Platform
2026-05-26 16:50:36 +08:00

16 lines
543 B
JSON

{
"achieved_tflops_per_gpu": 0.003108013341562435,
"achieved_tflops_per_gpu_theoretical": 599.6591076197008,
"epoch": 7.0,
"loss_nan_ranks": 0,
"loss_rank_avg": 0.2352302223443985,
"mfu_percent": 0.0002196475859761438,
"mfu_percent_theoretical": 42.37873552082691,
"total_flos": 1042355907002368.0,
"train_loss": 0.3237440354855496,
"train_runtime": 20961.0568,
"train_samples_per_second": 2.835,
"train_steps_per_second": 0.177,
"valid_targets_mean": 3882.4,
"valid_targets_min": 1848
}