Files
b1_top2_seq/train_results.json

16 lines
545 B
JSON
Raw Permalink Normal View History

{
"achieved_tflops_per_gpu": 0.003964633261549306,
"achieved_tflops_per_gpu_theoretical": 758.8906220723896,
"epoch": 7.0,
"loss_nan_ranks": 0,
"loss_rank_avg": 0.24622520804405212,
"mfu_percent": 0.0002801860962225658,
"mfu_percent_theoretical": 53.631846082854395,
"total_flos": 1226814568660992.0,
"train_loss": 0.3031141919652676,
"train_runtime": 19339.9756,
"train_samples_per_second": 3.578,
"train_steps_per_second": 0.224,
"valid_targets_mean": 2796.5,
"valid_targets_min": 1104
}