Files
Kimi-K2T-ling-coder-sft-san…/train_results.json

16 lines
542 B
JSON
Raw Normal View History

{
"achieved_tflops_per_gpu": 5.394061256937885,
"achieved_tflops_per_gpu_theoretical": 792.7696677006214,
"epoch": 7.0,
"loss_nan_ranks": 0,
"loss_rank_avg": 0.12738347053527832,
"mfu_percent": 0.5454055871524656,
"mfu_percent_theoretical": 80.15871260875848,
"total_flos": 6.410364364720701e+17,
"train_loss": 0.25545546521479906,
"train_runtime": 14855.1436,
"train_samples_per_second": 3.736,
"train_steps_per_second": 0.234,
"valid_targets_mean": 3825.4,
"valid_targets_min": 1564
}