Files
a1-stackexchange_tor/train_results.json

16 lines
548 B
JSON
Raw Permalink Normal View History

{
"achieved_tflops_per_gpu": 0.0026187013159839974,
"achieved_tflops_per_gpu_theoretical": 555.6163483029995,
"epoch": 7.0,
"loss_nan_ranks": 0,
"loss_rank_avg": 0.24078989028930664,
"mfu_percent": 0.00018506723081158992,
"mfu_percent_theoretical": 39.266173024946966,
"total_flos": 1125904534470656.0,
"train_loss": 0.35793851753530226,
"train_runtime": 26871.7295,
"train_samples_per_second": 2.627,
"train_steps_per_second": 0.164,
"valid_targets_mean": 2557.6,
"valid_targets_min": 1102
}