Files
exp-uns-tezos-80x_glm_4_7_t…/train_results.json

16 lines
543 B
JSON
Raw Permalink Normal View History

{
"achieved_tflops_per_gpu": 3.2265754742464843,
"achieved_tflops_per_gpu_theoretical": 304.9114106540851,
"epoch": 7.0,
"loss_nan_ranks": 0,
"loss_rank_avg": 0.14714080095291138,
"mfu_percent": 0.32624625624332504,
"mfu_percent_theoretical": 30.830274080291723,
"total_flos": 1.2559500582227804e+18,
"train_loss": 0.328229405380178,
"train_runtime": 48656.4652,
"train_samples_per_second": 1.448,
"train_steps_per_second": 0.09,
"valid_targets_mean": 2880.2,
"valid_targets_min": 1091
}